Watching companies go through the same PLG-to-enterprise growing pains over and over again makes me feel like I’m Bill Murray in Groundhog Day. I’m not really seeing this at Lovable and I will do everything in my power so that it will continue this way, but as I talk to other companies, it is always a painful reminder how wide-spread this is…

So this post isn’t really about the solutions. Okay, yeah, obviously I have strong opinions about how we can fix some of this stuff. But for now sometimes it’s helpful to just realize you’re not the only one having these discussions. The good news for you is… your problems aren’t so unique. Not such great news for our industry, though.

So I made a bingo card of my um, favorite conversations.

Sponsored by Fin AI: Join Pioneer on Oct 7 to explore how leaders from Anthropic, Clay, and Kalshi have transformed their customer experience and redefined what’s possible with AI. You’ll also hear from the co-founders of Fin on their vision for the future of CX, and be the first to see what’s coming next. Reserve your spot →

Sponsored by Chargebee: Earlier this year, I wrote: “evolve your monetization model - because I guarantee yours needs help.” Chargebee builds billing and monetization infra that powers just that evolution for fast-scaling, usage-native businesses. Learn more →

‘Why do we need 10,000 $10 customers when we could get one $100K customer?’

This is the original sin.

At some point (usually after the first few big enterprise deals close) someone does this math. And I admit that the math checks out. One customer paying one hundred thousand dollars is indeed the same amount of money as one thousand customers paying one hundred dollars. We have Excel. We have confirmed it.

The problem is that those two revenue streams don’t have the same acquisition motion. If our enterprise business grew out of our PLG base, that $100K customer may have started as a $25 self-serve customer. Someone discovered the product, started using it, invited their team, usage spread through the company, and eventually the account got to be large enough that security, procurement, IT, or Sales got involved. The $100K contract is the final result of a much larger system underneath.

But big ACVs are hard to resist. Especially once someone puts them in the board deck. And that’s when the conversation turns into: why are we spending all this time optimizing self-serve when Enterprise is clearly the bigger opportunity? Let’s hire enterprise sellers. Let’s build an outbound motion. Let’s move the roadmap upmarket. Let’s focus Marketing on executives. Look at how much more efficient this will be.

Except that means changing literally the entire acquisition model. That’s going from capturing enterprise demand generated by your product to trying to somehow generate enterprise demand from scratch. Those are not the same motions. They require different channels, different talent, and different economics. And a massive new marketing and sales budget.

And while everyone is busy building the shiny new enterprise machine, the PLG engine is dying a slow, sad death. Acquisition slows. Activation gets worse. Fewer teams adopt. Fewer accounts naturally expand into those beautiful $100K contracts. Then, a year later, everyone gets together to figure out why enterprise pipeline is drying up.

‘Let’s count every self-serve signup as a lead!’

‘Hm, pipeline is slow? Well, PLG has lots of signups. Therefore, signups = leads. Look at us aligning our go-to-market motions!’

‘Great, now let’s call everyone 10 minutes after they sign up.’

Nothing says ‘product-led’ like signing up for a product, clicking around for 7 minutes, and immediately getting a phone call from Gary asking about your enterprise budget.

But why stop there? Let’s put a mandatory sales form directly into onboarding: Company size. Role. Phone number. What are you hoping to accomplish? Would you like to talk to Sales? Please tell us your annual budget before we allow you to discover whether our product even works.

This works in the short term, since your ‘lead’ volume will explode. But… wait a minute?? Why are all these leads converting so poorly!?

‘These product leads are garbage.’

Yes. Correct.

We decided to call everyone a lead.

We took people who were trying to use the product, interrupted them before they experienced any value, collected their contact information, routed them to Sales, and then discovered that many of them weren’t interested in talking to Sales.

This should not have been a surprising research finding.

You can’t ambush users and try to force/trick them into taking sales demos and call it product-led sales. A good PLS motion uses product behavior to identify when an account is developing a problem that Sales can actually help solve. Watch for things like team adoption, usage expansion, governance needs, plan limits, multiple departments showing up, and whatever signals actually correlate with an enterprise opportunity in your business.

Otherwise it’s just harassing people. Plus the added side-benefit that these people may be our most loyal community members and power-users, who were huge advocates for us… until they started getting bombarded. Whoops!

‘We’re losing sales deals to self-serve.’

This is always my favorite because you have to admire the mental gymnastics it takes to call this a loss. We acquired a customer… they liked our product enough to pay us… and they chose the buying experience they preferred. The only problem is that they didn’t talk to the human we wanted them to talk to, so somehow we have decided this is bad news.

I know why this happens. Sales teams have quotas, companies have targets, and somewhere there is a very beautiful forecast with a very specific number in a very specific cell. But the customer does not give a shit about our org chart. If they can solve their problem with self-serve today, let them. Maybe they expand tomorrow. Maybe they need enterprise 6 months from now. Maybe they never do. The goal is to let customers transact in the way that makes sense for them, not force them into the channel that makes our internal attribution model look nicest.

And once we start competing internally over who gets credit for revenue, things get weird very quickly. We’re no longer competing with our competitors. We’re competing with ourselves. Amazing strategy.

‘Let’s move this self-serve account to Enterprise.’

If you can’t beat ‘em, make ‘em join you, right?

Nothing actually changes for the customer - same product, same usage, same people, same revenue. But that doesn’t help us hit our enterprise revenue growth goals, now does it?

So we take $50K of self-serve revenue, move it into the Enterprise column, and congratulations, Enterprise grew by $50K. We did not acquire a new customer. We did not expand the customer. We did not create $1 of incremental revenue. We just copy-pasted a number into a different column on the revenue spreadsheet and now we’re all getting bonuses!

Unfortunately, we may have actually made the revenue less valuable. The customer who previously self-served now gets a salesperson, a CSM, higher-touch support, contracting, account management, and a bunch of other enterprise services that aren’t cheap to deliver. We successfully took revenue we already had and made it more expensive.

This is usually an incentives problem. If Sales gets quota credit for moving existing self-serve revenue into Enterprise, you should expect Sales to get extremely talented at moving self-serve revenue into Enterprise. People tend to get very good at the thing you pay them to do.

One simple way to avoid this is an ARPA multiplier for channel changes. If a $10K self-serve account converts to a sales-owned account, that’s not a $10K Enterprise win. Just create a policy that requires an account to become much larger (like, 2x or 3x its existing ARPA) before the migration gets treated as successful sales expansion. You should pick the multiplier based on your own economics, but you need to create incremental value rather than just reclassifying existing value.

Otherwise you’re measuring revenue migration, not enterprise growth.

‘Can we move this feature from self-serve to enterprise? It will help us close more deals.’

Y’know what’s even better than moving happy self-serve customers over to enterprise plans? Stripping happy self-serve customers of their core features to ‘encourage’ them to buy a higher tier plan.

This one is especially fun when the feature already exists. Like, customers are already using it. It may even be one of the reasons they love the product. But now we’ve discovered that it could help Sales close deals, so naturally the solution is to take it away.

The logic sounds something like: enterprise needs stronger differentiation, this is a valuable feature, so moving it upmarket gives Sales something to sell. Except that this is removing value, not creating it.

There are absolutely features that belong in enterprise. But your default instinct should be to push product value down, not constantly pull it up. The goal for packaging is not to make self-serve shitty enough that Enterprise finally looks compelling.

‘Let’s have all new features go into Enterprise plan. Enterprise needs more differentiation.’

This is the more sophisticated version of the same idea. Instead of debating individual features, we’ve created a policy: innovation belongs to Enterprise unless proven otherwise.

Which is a fascinating strategy for a company whose growth was built on people being able to discover, try, adopt, and fall in love with the product without talking to anyone.

Enterprise differentiation should come from solving enterprise problems, not simply getting first dibs on everything Product builds. Otherwise your self-serve product slowly becomes a museum of everything you built 2 years ago, while all new value gets locked upstairs.

And then 6 quarters later, someone will of course ask why self-serve growth has slowed down.

We should probably launch an investigation.

‘You have to build this. I already sold it to the customer.’

I love this version of roadmap planning.

Yeah, of course Sales should be bringing customer needs into Product - they’re on the front lines and have tons of context about what customers want, what’s blocking deals, and where the product is struggling.

But ‘customers keep asking for this’ and ‘I promised this to a customer and told them it would ship in Q3’ are not the same thing. That’s moving from a mutual conversation to a hostage situation.

And the opportunity cost is almost impossible to see. The team is all clapping for the $300K deal marching its way through the Salesforce dashboard, but it’s impossible to see all the thousands of future customers who won’t activate, retain, or expand because Product spent the quarter building something that one enterprise prospect ‘needed.’ Actually, no one will ever see those customers at all, because they’ll never sign up.

‘You can’t ship that!’

Yep, opposite side of the same coin. Build the stuff my clients say they need, but also for the love of all that is good, please do not release new capabilities that would disrupt the sales cycle. Or without giving enterprises clients 3 month notice( nothing screams innovation more like 3 month notice).

This may include but is not limited to:

• New features that the sales team has not yet received enablement training on

• New functionality that improves efficiency and might change the original value prop

• New navigation, which could possibly require new customer education materials, client exec buy-in, and an entirely renegotiated MSA and SOW.

This is how a team that was once famous for shipping super fast and resolving customer concerns within days becomes an org that for some reason can’t even patch their own bugs without 7 layers of approval and 3 month notice period.

How’d you do?

Maybe checking off all those Bingo boxes doesn’t really feel like winning, but at least you know you’re not the only one playing. Do share in the comments if you have any ‘must do’ ways of solving these ‘problems’.

Edited by Jonathan Yagel.

Yep, Lovable is a PLG company. Most of our revenue comes in self-serve. We also have an amazing sales team, but our main motion is still very bottom-up. Prosumers, founders, people building personal projects, and a surprisingly large number of people using Lovable inside companies are still just swiping their credit card.

And yeah, we’re an AI product, so every time someone uses Lovable, it costs us actual money. But a huge part of our growth strategy is still based on giving the product away for free.

I know. Somewhere, a finance person just felt a disturbance in the force. Because AI made everyone panic about gross margin, and in response a bunch of companies are hiding the exact thing customers need to experience before they’ll pay.

Sponsors: Elena’s Picks

You don’t want generic ads, right? To make things more valuable for you, I’m featuring products and services that I personally recommend or am trying out:

→ Research Flow by Typeform: I built and ran an entire qualitative study of 300+ people (for my ‘What people really think of IC roles’ report) with the Research Flow product and got to work closely with the Typeform team to hear their vision for it. I’m excited for all teams (growth, marketing, product, everybody) to get to do way more high-quality qual research.

🎁 Offer: Schedule a demo and get $500 off. See terms and conditions.1

→ TrustedRouter: I’m exploring this one because it feels like my LLM subscriptions are multiplying faster than my streaming subscriptions. If you’re building something and need multiple models, you can use TrustedRouter to get all of them (600+!) in one place. They automatically handle switching, in case one of the services goes down or becomes cheaper. I also love that they take privacy protection seriously - they use cryptographic proof to show they never see/store your prompts.

👀 Learn more. Bonus: No subscription fees! They just charge by usage.

→ Schematic: Anyone else figured out how to do non-seat-based pricing? It’s messy. So I’ve been checking out Schematic’s new usage-based billing platform. Ideal for working with enterprise customers who want to know their usage in real time, plus runtime entitlements, access control, metering, ramp periods, and all the other stuff you need if you’re not just gonna charge a flat rate every month.

👀 Sign up for free if you want to explore shifting to usage-based billing. Or, check out their AI Credits & Metering virtual workshop next month.

AI makes this even more important

The way we all think about product giveaways needs a refresh because the OG strategies were built during the SaaS era, and cloud-based software is unique in being super cheap to scale. We were all happily running around with 80-90% gross margins. Giving away extra seats (which usually cost, like, pennies in server costs?) wasn’t exactly going to send the CFO into cardiac arrest.

But then along came AI, which started destroying those beautiful margins. Depending on the product and model, you could have 0 margin, or even have negative margins. Even after optimizing things, you’re doing an amazing job if you’re getting over 40%. So of course companies responded by trying to protect those margins. AI functionality went into the highest-priced plan, or became an add-on, or (my personal fav) got locked inside Enterprise where majority of your user base could admire it from afar.

Because how dare you give something away when it costs money? Math needs to math!

This is one of the biggest monetization mistakes companies are making with AI right now.

Why? Well, for one, you’re literally blocking users from experiencing your product’s main value.

AI is still pretty new, so most people don’t really get what it can do for them until they use it. Even if someone describes it (and I know your marketing team is working overtime here), it’s super hard to understand the value until we experience it ourselves.

(Also, don’t you think all most of the marketing is kinda saying the same thing? Some version of “built for the AI era,” “AI-native,” “agentic,” “supercharge,” “at scale,” “focus on what matters,” and anything involving “redefining what’s possible.” Meh)

But please remember, your users are making a simple calculation: The perceived value needs to be higher than the price you’re asking them to pay. And by the way, that ‘price’ isn’t just dollars. It’s also the time and effort it takes to set the thing up, learn it, convince coworkers to use it, and a million other things that are usually required to get to an actually get value out of a product.

Putting AI behind a monetization wall before they’ve “redefined what’s possible” is asking them to pay for something they don’t understand yet.

Think of it as marketing spend

When your finance team hears about wanting to give away AI product usage away, they immediately see it as a giant cost center that’s going to destroy your margins.

Our whole point at Lovable is to reverse this:

I think about the cost of giving away free product usage as part of our acquisition spend. Which means it has to compete with every other place we could spend our marketing dollars.

We generally apply a 3-month payback threshold to our free acquisition strategies. If I give someone $X worth of Lovable, how quickly do I make that money back based on conversion into paid accounts, retention, renewal, and expansion?

Then I compare that against other acquisition channels.

Most people think the 2 options are: ‘give the product away vs. keep the money.’ But the actual alternative is usually ‘give the product to prospective customers vs. spend multiple times more money trying to acquire those same customers through traditional marketing channels.’

I’d rather make our users richer than Google. (No offense, Google, we still spend plenty of money with you.) But if our product can acquire customers itself and sell itself, we would obviously rather invest there and create more competitively defensible growth machine.

Plus, this approach gives you a chance to engage and monetize customers even if it doesn’t work out the first time.

Think about all the users you have that tried your product, but didn’t fully convert. With this approach, at least they’re in your system and you have their contact info, so you have a chance of resurrecting them. If they’re coming in through paid ads, you only get one shot. If they don’t convert then, they’re never coming back (unless you pay the same amount of money in the future to hit them with ads again.

Also, remember: even companies selling physical goods with real costs always do trials, giveaways, and samples. All.The.Time. SaaS has just gotten spoiled rotten.

(Side note: I know some people will discount this, like, ‘Oh, you can only do this and say all of this because you’re VC-funded and Lovable just raised $400M.’ But I’d actually say that if you evaluate it very strictly from a payback perspective, every growth channel or tactic has a cost. Paid is obvious and finance may complain about giving away tokens, but even the ‘free’ growth activities have a cost because you still need to hire people to execute against them. So yeah, VC money means that we can do this a lot faster, but every company should consider this channel as a way to grow in a smart, efficient way.)

Show them the magic

I’ve written a lot about freemium before, so I’m not going to turn this into another freemium manifesto. The most important thing is simple: give away enough of your product for someone to reach the ‘oh shit, I can’t believe this is happening’ moment.

For us, that moment is the first generation - telling Lovable, ‘Build me a cool app that compares the sizes of different dinosaurs’ (or whatever) and watching it appear on your screen. We strive to make this experience free by giving users 5 daily credits (10 on the first day), capped at 30 per month.

That first generation is everything for us. We’re always working to make every generation great, duh, but that first one is super important because it’s when someone gets to experience translating something from their brain into working code on their screen.

And if your product requires multiple uses to really understand, make sure to factor that in. One great first session might not be enough. This is why our free credits refresh daily. Maybe you weren’t convinced enough to pay on day one, but tomorrow you can come back, make another edit, see a little more value, and build a habit.

I cannot take credit for this strategy because it existed before I joined Lovable, which is deeply inconvenient because it works extremely well.

Our free-to-paid conversion typically lands somewhere around 5-10%, and that conversion is what ultimately makes the payback math work.

But hey, just remember that your product has to really sell itself :)

Work with others to give your product away

Remember that freemium is an entire system, not just a price point. To make this work, we’ve been offering free product usage through a bunch of different programs.

1. Partner giveaways - borrow someone else’s audience

I’d actually never done it at this scale before, but it’s worked surprisingly well for Lovable.

The premise is very simple: find another company or creator whose audience overlaps strongly with your ICP, and give their audience your product for free.

For example, you can get some free Lovable through LinkedIn Premium Perks. We have offers through Lenny’s Newsletter, Revolut, this beautiful newsletter you’re reading right now, and many others.

The partner gets value from being able to give away your product, and you don’t have to compete in the auction for attention that exists on Google, Meta, LinkedIn, and basically any ad platform.

Plus, your distribution is more efficient because you already know you have the right audience. If the partner has an extremely strong ICP match, we can get pretty aggressive with the offer - several months or even a year. That sounds insane until you look at the conversion rates into paid accounts, which can be up to 40-60%.

Again: the math needs to math. After viewing this through a payback lens, I am now a very big believer in this channel.

(Emily Kramer wrote a great guest post for Lenny about this last year: Ecosystem is the next big growth channel.)

2. Community and events - let other people activate your users

Our Lovable hackathons have been going gangbusters and are making me think that product-focused events are underrated. Not only do the champions and organizers get to be in the spotlight, but everyone who attends is set up to have a perfect onboarding experience.

Just think about it: People are trying the product out with a specific time constraint, surrounded by other people using it and also people to answer questions or help them get unstuck. And now they know a bunch of people in their office (or wherever) who they can talk about their new project with. This is pretty much an ideal environment to get people to not only use the product once, but to get them to stick around and keep using it.

Oh, and this has been an amazing enterprise outreach program too. Instead of doing a demo (bleh), we help enterprises organize internal hackathons on Lovable so their teams can actually experience the magic.

3. Startup programs - get into the stack before there’s a stack

If you want startups to use your product, give it to ‘em for free.

I’ve seen this work across many companies and categories. MongoDB has done it. I’ve heard similar examples from MathWorks. At Amplitude, our startup program was a major driver of sales pipeline, not just self-serve revenue.

I wrote a whole article busting myths about this approach, but the big one is approaching this as a tactic to drive short-term revenue. You can’t expect these companies to start really paying off right away, but it’s a great opportunity to get your product built into their core infrastructure, so that your contract with them will grow as they do. If you’re willing to wait, that’s a lot more efficient than trying to break into a hot company with an enterprise sales motion.

The post-graduation conversion I’ve seen from strong startup programs can be in the 40-60% range. If your product becomes more valuable as the company grows, I don’t understand why you wouldn’t want to get there early.

Before you start throwing free product everywhere...

There are a few guardrails because, unfortunately, ‘give away your product’ can very quickly become ‘accidentally sponsor the internet.’

First, measure payback. Giving away the product should not be treated as some vague brand investment where everyone nods thoughtfully and nobody knows whether it worked. Assign a dollar value to what you’re giving away, measure conversion, retention, renewal, and expansion, and compare the impact against your other acquisition channels. We use a 3-month payback threshold. Yours might be different, but you need one.

Second, don’t use generic coupon codes. They always leak or get shared to people who aren’t supposed to have them. They end up on coupon sites. Someone posts them somewhere they absolutely were not supposed to be posted, and then your carefully designed acquisition strategy turns into a charitable contribution to strangers on the internet. Always try to use unique single use redemption codes tied to individual users or eligibility.

Third, adjust your offers per market. Make sure you understand the economics in different regions, because they vary wildly. Like, if conversion or monetization changes a lot by country or region, you probably need geo restrictions. Again, the math needs to math. ‘Free’ doesn’t mean ‘we don’t care about unit economics anymore.’

And fourth, actually make it free. I’m a much bigger believer in giving someone a smaller amount of product usage for $0 than giving them more access for 20% off. Even if I can easily afford a few dollars, there is a big psychological difference between ‘try this’ and ‘buy this, but slightly cheaper.’ The second you ask for money, I need to make a purchase decision. When something is actually free, it removes that decision and lets me focus on whether your product is actually useful.

And yeah, people will steal from you 😠

Any valuable offer is gonna attract fraud. It sucks, but it always happens. There’s no magical strategy where you give stuff away on the internet and everyone politely takes only what they’re supposed to.

You need someone responsible for this.

At Lovable, our team builds systems to detect fraud, reclaim credits, and ban bad actors. These systems can catch and fix a lot, but you still need to assume that someone, somewhere, is currently trying to figure out how to exploit whatever generous thing you just launched.

I learned this personally through this newsletter. As some of you have seen, I started offering free Lovable credits to every paid subscriber, which I thought was a nice thing to do. But then…. one dude found a loophole, generated 600+ coupons, and started reselling them.

When I confronted him, I was expecting at least some version of ‘Oh shit, sorry.’

Nope.

Apparently stealing 600 coupons and reselling them was simply another entrepreneurial path I had failed to adequately appreciate.

This is annoying. It makes everything more expensive and more difficult for everybody. So, yeah, you’ll need fraud detection, eligibility rules, unique codes, monitoring, and remediation. And people will still occasionally find a way around all of it.

But even with all of that, it’s still definitely worth it.

Freemium forever

If you’ve done PLG in the past and someone (probably in finance) convinced you that AI killed freemium, go tell ‘em they were wrong.

Yeah, the stakes are higher because giving away product usage has an obvious and immediate cost to it. But if you want to invest in growth, I’ve seen for a fact that the cost of giving those tokens away can be much more efficient at driving customers and conversions and revenue than the cost of the more obvious marketing channels.

And as a bonus: You’ll have a big advantage if you can get your act together on this, quick. Right now, in this world where there’s very little feature differentiation between products, all the marketing sounds the same, and everyone’s saying the same thing. There’s so much damn AI confidence theater. So you’re gonna need to win customers over with your experience… and this whole approach works because it’s stripping out all the obstacles to getting people into your product experience, asap. If you’re looking to acquire customers in this extremely competitive market, give them your product and let them try it for free.

Edited by Jonathan Yagel.

So I decided to dig a bit deeper. I ran a study!

Some of my favorite highlights:

• Yep, people want to do IC work: 71% of people (including managers!) want at least half of their time to be spent on IC work. 27% said they only want to do IC work.

• Manager-turned-ICs wouldn’t trade it for anything: 75%+ of this group said they wouldn’t trade roles with anyone in the org. (vs. ~30% of current managers.)

• The IC cost is… influence: Yep, 32% mentioned comp. But 66% of people said the main downside of being an IC is lack of influence.

But the numbers themselves weren’t the main thing. This was n…

Read more

Read more

Meanwhile, I’m over here asking ChatGPT to rewrite the same paragraph for the third time because it keeps defaulting me into ‘LinkedIn wisdom post’ mode. GARH.

Look: I work at an AI company. I use AI all day. I strategize with it. I build products with it. I write with it. I analyze data with it. But even I feel like everyone else has cracked the AI code while I’m still using some basic tutorial version over here.

So I started asking a simple question whenever someone told me AI had changed their life: ‘Cool. Show me.’

And most of the time I see some basic workflows. Summarizing Slack. Answering emails. Doing scheduled scans. Performing research and booking something. Sending emails out of Claude. Okay…. it’s useful. But show me something that has become so critical in your day that if I took it away tomorrow, your work would actually fall apart. Show me something truly life changing.

It turns out that’s a super short list.

But the noise continues on volume 11/10.

So this is my desperate plea... Can we PUHLEASE stop this AI Confidence Theater, people? It’s doing more harm than good.

This post is sponsored by Firecrawl. Firecrawl is the web data API to search, scrape, and interact with the web at scale. Turn the live web into clean, structured data your agents can actually use.

(P.S. I personally use Firecrawl in my Lovable apps all.the.time - most recently to scrape this very blog so my AI double could use my latest posts as context.)

It is screwing a lot of stuff up

This fake AI hype is hurting all of us. This is not a victimless crime.

The reality distortion hurts innovation

The AI headline chasers and fakers are robbing people of a real, cool Ah-ha moment with AI tools that could get them excited and bought-in. The problem is when you see someone talk about their ‘life-changing’ (it’s always ‘life-changing’) new AI-powered system but… it doesn’t actually work. So you just assume that all of the AI stuff is BS and keep your head in the sand.

And instead of celebrating the saved time and reduced annoyance that AI actually can do right now, people are out here promising employee-replacing super agents… despite the fact that their agents only trigger 50% of the time and only deliver decent outputs if hand-fed very specific context (and get shit wrong most of the time)

This weird reverse-hustle hustle culture?

I didn’t have ‘Kind of misses toxic hustle culture’ on my 2026 bingo card and yet, here we are. Because now we have an even more toxic environment.

Five years ago, people wanted you to think they worked harder than everyone else. Today they want you to think AI does it all.

Five years ago, people shared and bragged about big outcomes: revenue, user milestones, fundraising, headcount. Today people are just highlighting how many tokens they’re burning. Everyone wants you to believe they have an invisible team of AI employees running their business while they think big thoughts. But where are those business outcomes, tho?!?

Five years ago, people bragged about waking up at 5am, doing cold plunges, having inbox zero, and working while endlessly Peletoning. Now it’s all about having 17 agents, not opening email, and AI running your business. It’s the exact same performance, just with different props.

Hiring for real expertise is a mess

This has also blown up the hiring process. AI has made average intelligence incredibly cheap and it has given everyone the vocabulary of expertise. It used to be that if someone started talking about vector databases, MCP, agents, memory, or RAG you assumed they knew what they were doing. Today, everyone knows the words. AI regurgitates some smart one-liner for them and 3 hot takes. Everyone sounds like they have an experience.

But sounding competent and being competent are completely different things. So how do you hire in the middle of that? Not that the old interview and hiring process was so great - I’ve been frustrated with it for years. But verbal interviews for real don’t work anymore. Just because you can ‘confidently’ explain MCP doesn’t mean you can build one or you’ve built a single workflow that anyone actually depends on. Case study and work trials become absolutely necessary to cut through the bullshit.

The FOMO is exhausting

All this leaves people feeling demoralized. Instead of getting excited about how these tools could help, it mostly just feels like we are all terribly behind and it’s all completely impossible.

Overall, our brains just can’t comprehend this level of change. Especially for people in tech, this is all happening so fast. Think about the fact that 200 years ago, we were just getting basic sanitation figured out, and now we have to figure out how to manage all-powerful (but also really inconsistent and sometimes disappointing!?) artificial intelligence that is going to help/replace our jobs? It’s just so much. We don’t need people yelling at us all the time about how we’re behind.

What I don’t like about it is that it creates a fake baseline. If everyone around you appears to have figured something out about AI that has transformed their work, then using AI to summarize meetings suddenly feels embarrassingly basic. You stop appreciating the things that genuinely save you time because someone else claims they’ve done so much more.

Why is this happening?

There are a few big reasons that come together to make this happen: social media rewards superficial hype, the stuff AI can do is evolving faster than we can keep up, and companies are incentivizing this sort of behavior all up and down the org chart.

Problem 1: Hype gets clicks

We live in a world where attention is the currency. And that attention is generally measured in impressions. This is the dopamine hit that people are chasing - can you go viral? Can you get everyone to engage?

The way to capture attention is to say something shocking or unexpected, so everything gets exaggerated. Legitimately cool use cases get elevated from ‘This saves me 15 annoying minutes every week’ to ‘ZOMG THIS CHANGED MY LIFE’ or ‘USE THIS WORKFLOW OR YOU’LL BE PERMANENTLY UNEMPLOYED.’

Problem 2: It’s the wild west out here

When we’re talking about content creators sharing their magical use cases on LinkedIn and X, the tricky thing is that AI really CAN do crazy stuff. If it were all fake, then we could all just ignore it. But people actually are building stuff all the time that could totally change how you and I work every day. So, we can’t just ignore it.

But it’s also hard to confirm. Like, are they using a different LLM model than me? Is my data not structured right? Did I mess something up in setting up the workflow? Or… did they just fake it, and all of this is impossible? There’s no way to know.

This is not a problem that we had to deal with previous phases of tech. No one was creating ‘I just built this INSANE setup with Dropbox’ posts to try to go viral… everyone would have just looked at the description and immediately called bulllllshit.

Problem 3: Marketing teams are selling certainty where none exists

Traditional software was ‘easy’ to message. Click the button, get the outcome. AI doesn’t work like that. It opens up more possibilities, use cases become loosely defined, and it needs the right context to succeed. Sometimes it’s incredible. Sometimes it’s just stupid or confidently wrong.

But read the landing pages and you’d think every AI product is a magical employee that never sleeps, never makes mistakes, and runs your company better than anyone else.

The thing about all this AI innovation is that it’s also the era of trust-based growth. If you over-sell, it will be discovered and no one will trust you anymore. Right now, maybe all of this hype feels like free attention and money. But if you take that route, you’ll pay for it in the end.

P.S. I don’t even blame marketing teams. They’re doing what marketing has always done: selling the best possible version of the product. The problem is that with AI, the gap between the demo and day-to-day reality can be enormous. And every exaggerated promise makes people trust the next AI product a little less.

Problem 4: Pressure from the top

Inside companies, this starts at the very top. Even above the execs - at VC level. Because the investors have been promised AI-powered miracles. And this means that execs need to demand AI-powered miracles from their employees. Which means that employees are being asked to do the impossible with (often underperforming) AI tools. Which means that employees are massively incentivized to perform the AI confidence theater to justify career progression, or maybe even protect their jobs.

Cut it out. You’re only hurting yourself.

If we’re going to stop the charade, it’s gonna take a group effort.

For content: Stop sharing (and spreading) the hype.

If you’re reading this and you’re guilty of posting exaggerated nonsense about your 100x AI Agent Workflows… stop it.

The good news is that there’s plenty of amazing stuff to be done. You don’t need to make things up or exaggerate. Just go deep on the stuff you know, and spend the time to build something really cool that actually can make real impact. If you share something that seriously changes how people work, they’ll always remember you. Some clickbaity title and false promises makes you forgettable or hateable.

And if you’re watching that kind of hyped content: You can be part of the solution, too. Hold your favorite creators accountable! Ask them to show you the receipts! If you know that something doesn’t work, don’t just let it slide.

And do NOT amplify it unless you try it yourself. If it doesn’t work and you spread it, you’re responsible, too.

For companies: Build your foundations.

The corporate side of this is a little bit trickier. Because again, it starts at the top.

• Execs: Set reasonable expectations. This means managing your investors to tell them what you can actually do. This also means setting benchmarks for your organization - creating unreasonable expectations only comes back to burn you.

• Managers / Team Leads: Focus on outcomes. Show your leadership what your team can move and use AI to amplify that. Don’t just start with AI and shrug your way to how it connects to business impact. If your team is still experimenting with AI tools and workflows, give them the space to build something that will really last.

• Employees: Figure out the actual impact. Ask for the time and training to get better with the AI tools that interest you! Everyone is talking about AI, but if you can actually connect it to fixing core business problems, you’ll be a real hero. Figure out how to anchor in outcomes! (Understanding and delivering this kind of impact is how you become a Hi-C, btw.)

My biggest career advice for anyone right now: spend a few hours a week trying something new with AI. It’s not a lunch, 5 mins between meetings, after work, or weekend activity. Learning AI IS the job right now. Make legitimate space for it.

Let’s just be honest & talk about the hard part

Seriously, people. Content creators. Company leaders. Employees. Marketing teams. Let’s just be honest about what we and our products can actually do with AI.

The first prompt is the fun part. The next thousand prompts is where real value is created. Let’s make sure we cover THAT part - what happens when the models change, the prompts stop working, the API behavior shifts, or the integration breaks and no one notices.

AI systems aren’t ‘set it and forget it.’ They’re living systems that need monitoring, evaluation, iteration, and constant tuning. That’s where the real work happens. That’s where the real value gets created. And somehow, that’s the part nobody posts about.

If you’ve built something amazing, tell people. But if you’ve built a promising prototype, say just that. We will still be excited, promise. If your workflow saves you 20 minutes a day, celebrate that. If your AI agent still needs babysitting or constantly messes up, admit it.

AI is already the coolest tech innovation of our lives… we don’t need to pretend it’s magical too.

Edited by Jonathan Yagel.

And maybe that’s not such a bad thing. The old way of experimenting had its limits:

• Focus on minor tweaks instead of big bets

• Optimizing superficial surfaces instead of foundational complexity

• Avoiding monetization (‘It takes 6 months of decision approvals & build out. If you are lucky.’)

• Product mostly testing through previous/post, instead of actual A/B tests or feature flags

• Focusing on fast wins - results measured in 2 weeks

Some of this made sense, back when product development was slow. There was so much bloat and so many surfaces that just creating awareness of features and explaining how to use them was a win.

But even then, many of those experiments were just pulling up revenue (if anything). They weren’t actually having incremental progress. But by improving the experience a bit or creating some urgency, we made this thing (which would have happened anyway) …

Read more

For decades, creating software required access to a small, highly specialized group of people concentrated in a handful of tech hubs and funded by venture capital. That meant only ideas large enough to justify those costs ever made it into the world. As the cost and complexity of building collapse, that constraint disappears. The winners won’t just be faster developers. They’ll be teachers, real estate agents, accountants, coaches, consultants, small business owners, and domain experts of every kind solving problems they understand deeply. The future of software won’t be defined by a tiny group of elite builders. The future of software will be defined by everyone else.

Get ready to start seeing (and hopefully building) more and more p…

Read more

Sure, having the right AI tools and skills is part of the equation. But even the best employees with all the most magical tools will get trapped… if your org structure and culture don’t adjust.

Right now, many companies already have the technology they need to go much faster. The blocker is company systems that are mostly designed to prevent things from happening. The power is centralized and all the team members are treated like a risk vector. Exhausting approval cycles, super tight boundaries on roles, unbreakable title-based hierarchies, and a whole tier of middle managers whose main job is to keep everyone in line.

All of this pretty much screams at employees, ‘We don’t trust you, stay in your lane.’

And it should be obvious: Rapid innovation cannot happen in that kind of environment. Super high-impact individuals and high-velocity teams will shrivel up and die.

Everyone’s talking about AI agents, but what you really need is employees with agency. And if you don’t figure out how to give them the access and autonomy to make decisions on behalf of your organization, all of your AI innovation efforts are screwed.

Sponsored by Optimizely’s Test + Learn Event (with me!!): On June 17, I’ll be speaking at this free, virtual event about how testing has shifted:
- Acquisition testing → Monetization
- Tiny conversion lifts → Bigger strategic swings
- “Growth at any cost” → Growth with margin discipline
- Endless optimization → Faster learning cycles
→ Register here.

Sponsored by Amplitude: Build better products with our AI analytics platform. With powerful AI Agents embedded across our platform, teams can analyze, test, and optimize user experiences faster than ever. → Learn more

‘Protecting’ employees by limiting access and autonomy

Most modern companies are pretty much built around a command-and-control, power-at-the-top model. Each employee is treated like an assembly-line worker with one specific function. ‘Do your job, don’t ask questions, we’ll do the thinking for you.’ The goal is to bring order to the chaos of big, broad organizations with lots of moving parts. If everyone thinks for themselves… who knows what could happen!

There are a couple of obvious ways that this gets applied:

1 - Access is gated by titles

One of the big reasons everyone is striving for bigger titles and team management positions is… that’s where you’ll be heard. That’s where you can actually get the required information. That’s where you get invited to the right meetings and channels.

Companies say they want high-impact ICs, then hide the information required to have impact behind manager titles.

Literally. When you get a manager promotion, you’re in a separate Slack channel where information is being shared with you that you had no access to, before.

We make assumptions that employees are not to be trusted unless they’ve earned their title - even if that was external, earned at another company.

2 - Only the bosses get to make decisions

Decisions are suuuper costly in organizations. It takes forever to make them and it’s almost impossible to reverse them, so you better get them right. Which means that everything is routed to the Very Important People who can make those decisions.

So decisions are slow because they’re expensive. They’re expensive because they’re slow.

In some ways, the whole system is trying to ‘protect’ the employees - because to be fair, the employees don’t have all the information that the managers do, and it would be easy to make the wrong call without the appropriate context.

The virtuous cycle of faster, cheaper decisions

The AI-native orgs strip out the status theater around the work. Anthropic is a good example: the visible structure is basically “member of technical staff” or “member of non-technical staff.” Not because titles are evil, but because titles often become permission structures. They decide who gets context, who gets heard, and who gets to make decisions. And that is exactly the kind of drag AI-native companies are trying to avoid. Which leads to:

A - Faster decisions.

If you ungate access to information and encourage autonomy, suddenly… the context doesn’t have to pass up-and-down the chain for every little thing.

What’s funny is, lots of companies are realizing how important it is for their AI programs to have sufficient context. ‘If the AI doesn’t have context, it can’t do a good job!’

Well, guess what? Employees also need context to make the right decisions and move forward quickly. Otherwise (just like an AI), they’ll make the wrong decisions or get stuck.

B - Cheaper decisions.

And when you make adjustments on the same day (rather than waiting until the next weekly/monthly/quarterly ‘strategy’ meeting or exec offsite), the cost of a wrong decision will also fall off. In the previous world, the cost of a bad choice was astronomical. It would take so much time to get out of it. But if everybody can run at a really quick pace, it’s okay to make a wrong turn! You can go a bit down that route, turn around, and be back on the right course, faster than other teams even start moving.

Management Bloat

In the traditional org structure, middle management is a feature, not a bug.

The primary job for this role is to travel decisions and information up and down the chain. Up and down, down and up, over and over. It’s a big game of telephone.

This is intentional. Companies don’t trust the junior employees with sensitive information or believe that they even need it.

Now, I will say that managers also do other things… but the structural need for those is being reduced, too.

• Manage cross-functional gridlock and unblocking: ‘Who’s going to do what’ and ‘Where can my team’s priorities fit into your team’s calendar’ and ‘When are we going to find time to support your initiatives,’ etc. But fewer teams and more centralized information make this less necessary.

• Career coaching and guidance: Always important, always necessary… but is the manager really the best career coach for all of their reports? From a personality fit and skills fit perspective, maybe not. This can be more democratized, too.

Companies need to realize that pushing AI tooling into their teams is not going to automatically fix the velocity problems. Yes, having those increased capabilities can empower them, but if they’re still blocked off by bureaucracy and their own limited context, they’ll never move fast.

How do you make this shift?

Okay, company leader. What do you do about this? You’re not going to fix it with a new AI tool, an all-hands announcement, or a slide that says “move faster.” This stuff is too baked into how companies make decisions.

And to be 100% honest - I’m not sure how you switch it. I can speak to the differences I’ve seen in how Lovable operates vs. lots of other companies I’ve worked with, but I don’t know how to make the shift.

But here’s my hypothesis: Don’t start by trying to rewrite the whole org. Prototype it. Try it out with some R&D or innovation teams to see what they can do. Give them aggressive targets and a flat structure, so that their full team has the autonomy and access to actually get things done. Put your sharpest, most high-agency people in the group to start and see how they do.

And be honest: you may need some new people in that group. Not because the current team is bad, but because organizational muscle memory is real. If people have spent years being trained to wait for permission, it’s hard to suddenly tell them, “Actually, now you’re allowed to act.”

The weight of responsibility

Employees need to listen up to this, too. Because of the design of many organizations, lots of employees haven’t experienced the downside of autonomy. Yeah, it sucks to just be treated like an assembly-line robot, being told exactly what to do… but that also spares you from the legitimately terrifying possibility of something going wrong and it really, truly being your fault.

At the end of the day, that’s why leaders get paid more: Accountability. They’re the one who is responsible for a thing getting done. If there’s no one telling you what to do, you have to figure it out for yourself. And if your launch flops or conversion craters or something else vital to the business falls apart… that’s on you.

Fortunately, with the right culture and a high-velocity environment, everything is figure-out-able. But as more companies figure this out and start handing out greater context and autonomy, people will need to consider whether they actually want it. High-impact roles come with status and comp, but there’s always a cost.

Agency > Agents

If your company is anywhere near tech, everyone is probably talking about agents. Are AI agents going to replace all employees?

I don’t think so.

Although I definitely believe that everyone’s current jobs are changing faster than we want to admit, it’s not agents we need to worry about.

At Lovable, we use AI agents internally to make departmental context easier for anyone to access. Each agent has a “parent,” usually the person with the deepest context in that area, who is responsible for keeping the agent accurate and up to date. That matters because it removes one of the biggest excuses for centralized decision-making: “People don’t have enough context.”

But the agent is not the unlock. The unlock is what happens when employees can actually act on that context. That’s the irony of the “AI agents will replace employees” conversation. Agents don’t have agency. They wait to be told what to do. High-agency employees do the opposite: they find the signal, make the call, and push the work forward.

So yes, build agents. Assign them parents. Keep them useful. But don’t confuse the tool with the operating model. The goal is not more agents. It’s more agency.

Trust-based growth starts with Trust-based teams.

A few months ago, I told you that Growth is now a trust problem. Software companies in particular are getting disrupted from multiple angles and the way to survive is to build a connection point with your current and prospective customers.

But that trust isn’t just something you have to build outside of your company. In fact, I don’t really know if it can exist outside of your company if it doesn’t exist inside of the company first. All of the techniques that make external trust-building possible start with you trusting your own employees.

Build in public? You’ve got to let people talk about what they’re actually doing. Sharing legit failures is scary, but that’s how you build credibility.

React quickly to user needs? You can’t have a centrally-managed, long-term roadmap that pushes urgent requests out by a quarter or two.

The old way of doing things was all built around maintaining a polished, fault-free exterior. Now, people want to see some humanity. Unique quirks and imperfections that show off an actual person on the other side of the screen. And that can’t be faked from the top. You’ve got to trust your employees enough to do that - and give them the access and autonomy to make it happen.

You can still adjust to the right internal and external approaches to make all of this work, but you’ve got to make that call, and quick.

How fast can your company make this decision?

Edited by Jonathan Yagel.

My answer is: It’s all about the team. And the team profile is not at all what you’d expect for a company that has $400M+ in ARR. It takes all types of people. Yes, we’re big on AI-nativeness, but that’s not all. You have to have the right blend.

So I’ve been telling people that you need farmers and you need cowboys, and you need old guard and you need new guard. And most importantly: All of these groups need to actually get along.

Read more

Edit: The original title of this post was ‘You’ll lose your job in 2027.’ I changed the headline.

A few readers shared that the original title felt painful in the context of recent layoffs. After reflecting on that feedback, I agreed the framing missed the mark. I’m sorry.

I still stand by the core message: our jobs are changing rapidly, and the point of this post was to encourage people to proactively adapt before change is forced upon them.

For anyone working in tech, I’m sorry to say this, but there’s a real chance you will not have the same job by the end of 2027.

Some of that will be because of layoffs. Intuit and Meta just announced cuts, and I don’t think anyone seriously believes we’re done seeing more.

But even if you keep your seat, your actual job is going to change. A lot. The responsibilities, expectations, tools, scope of what you’re capable of accomplishing, bar for what ‘good’ looks like, and number of people needed to do the work are all being rewritten every day.

It’s already happening to me. I know it will affect you too. In some cases, I fired myself out of parts of my own role. In other cases, I got fired out of responsibilities by the pace of the company, the shape of the org, and the reality of AI. Things I used to own either became less relevant, moved to someone else, got automated, or needed a totally different operating model.

So yes, I technically stayed at the same company. But I do not have the same job. And more likely than not, neither will you.

Change is coming. Which is great news.

Sponsored by Optimizely: Join the FREE 2-hour virtual event on June 17 for digital leaders. Discover how to run more experiments with AI, hear from Elena Verna (me!), see 5 agent demos end-to-end, learn how Salesforce, BBC & ASOS scale testing, plus exclusive content. → Save your spot

Sponsored by Amplitude: build better products with our AI analytics platform. With powerful AI Agents embedded across our platform, teams can analyze, test, and optimize user experiences faster than ever. → Learn more

Wait, great news?? Yes. Because right now, it’s not too late. You can start rebuilding yourself before someone else decides what version of your job survives. Imagine yourself like Scrooge waking up on Christmas morning, with a new lease on life.

You still have your job! You still have time to make changes! But… it is time to start preparing for those changes.

You get to decide what to do about it.

The first step is to be honest with yourself about this reality. Yes, it’s uncomfortable. Yes, it’s scary. But the only thing worse is pretending that it’s not happening. (see: Get your head out of the sand.)

But you better not stop there. Here’s what I would put on my ‘Prepare for a Job Change in 2027’ action list:

Step 1: Ask yourself: do you actually like your job?

This is uncomfortable, especially if you’ve been in it for a while. But as things keep changing, the competition is going to get more intense. You’re not going to win out in the same role over someone who loves what they do.

And if there are going to be dramatic changes in your role, maybe this is a chance to finally pivot into something you do get excited about. Maybe that’s a more niche speciality. Maybe it’s switching industries. Maybe that’s going back to IC work (see: IC work is the new career flex - which is what I did).

Step 2: Start exploring career optionality.

I’ve been beating this drum for years. If you’re ever going to do it… now is an ideal time to get started.

If you’ve been telling yourself, ‘I’m too risk averse to go out on my own. I need the stability of a full-time gig.’ Well, I’ve got good news and bad news for you.

Your current path is no longer a safe, stable option.

Even before all of this rapid AI transformation, the stability of full-time work has been pretty sus: If you’re junior, you’re the closest to being replaced. If you’re senior, you’re an expensive line item.

Think of all your friends and colleagues who have lost their jobs despite being awesome at what they do, and you have to acknowledge that expecting your career to last indefinitely is just staying on the Titanic while the lifeboats float away.

So get started right now. Like, today. Because the good and bad news is… solopreneurship doesn’t happen overnight. I wouldn’t recommend that anyone should leave their job with no momentum or progress. It usually takes a while to figure out what you want to do and what the market will pay for, so start that process right now.

Then, by the time you’re ready for a change - whether that’s because your job changes or you’ve got enough progress that you’re ready to go all-in - you’ll have the foundation in place.

And guess what, you’ve already got a great headstart. Just think of your career options as a product you are working on…

Optionality 1: Your skill IS the product.

I went down this road, and I’ve seen many other people go this way too. This path usually leads to a services business: consulting, advising, freelancing, or some other version of trading your time for money outside a full-time employment model.

I started investing in this path about five years ago. Today, I make more from solopreneurship than I do from the cash comp of my full-time role. To get started, think of your brain as a product you need to build a company/growth around:

• Acquisition: Where can you find new clients? Tell your network that you’re looking for 1-2 clients. Share your unique POV on LinkedIn or Substack and chat with people who seem interested. Counter-pitch those recruiters in your DMs! Post your Claude Skills (and Lovable skills, obvi) to show off your expertise.

• Activation & Retention: What’s the ‘Ah ha!’ moment you can deliver to your customer? What is your unique value proposition? Is it one-off or ongoing?

• Monetization: Charging an hourly rate is a good place to get started, but monthly retainers are better to protect your time.

I wrote a deep dive on this approach, here: Is solopreneurship right for you?

Optionality #2: Get yourself a Mom ‘n’ Pop SaaS.

It’s become insanely easy to actually build software products - are you taking advantage? You can take your specific expertise, package it into software for a niche audience, and sell it as a one-off purchase or a monthly/annual subscription?

I call this ‘Mom ’n’ Pop SaaS.’ Not every software business needs to become a venture-backed unicorn cosplay exercise. The goal can be much simpler: keep it small, useful, profitable, and yours. Maybe it only makes $3k/month. But $3k/month in semi-passive income is ahmazing. With the cost of development disappearing, this now has very positive ROI.

And lack of technical background is no longer your blocker - AI tools help you build the actual thing. Cough… give Lovable a try, cough.

Which brings me to the next must-have…

Step 3: Become AI native.

Most people hear the term ‘AI-native’ and either start cheering loudly or throwing up in their mouth a little bit. The reactions are extreme. But one thing is clear: getting good at this will help you keep your job, reshape your job, or find your next one.

Just please get going right now. And don’t overwhelm yourself - start small. The point is that AI-native is not about having an LLM open in your browser all the time. It’s about how you think about the value of your work and how you approach it.

At this point, I see AI-native-ness as a series of tiers, in terms of how you’re actually using LLMs:

Tier 1: Writing buddy

This is a great place to start. All of us write stuff… so use AI to improve the clarity of your written communication, because good communication is still somehow one of the most rare corporate superpowers. Pro-tip: Write your own version and then have AI critique it, so you actually learn. Don’t be sloppy!

Prompt examples:

• “Make this more clear”

• “Help me make my point”

• “Add more evidence to back this up”

• “Rewrite this so a skeptical exec would buy in”

Tier 2: Meeting partner

Granola is in every meeting with me. It writes my todos, reminds me what I committed to, and it legit gives me good manager feedback!

Prompt examples:

• “If you were my manager, what are my strengths and areas for improvement?”

• “Summarize the decisions and who owns what”

• “What did I miss that I should follow up on?”

• “Where did I talk too much or not enough?”

Tier 3: Thinking partner

LLMs are agreeable to a fault and will just ‘Great point!’ you to death, if you let them. But if you approach conversations from a different angle, AI can help challenge your logic, find holes in your thinking, pressure-test your strategy, and bring in additional research.

Prompt examples:

• “Disagree with me. What’s the strongest counter-argument?”

• “Give me 20 ideas, then cut to the 3 best”

• “What am I missing or being naive about here?”

• “Do some deep research”

Tier 4: Builder

Don’t just talk about it. Build it. This could be internal tools or external apps (Satellite Apps, anyone?). Or landing pages, full websites, prototypes… anything.

You should probably use Lovable for this.
Paid subscribers to Growth Scoop get it for a year for free!

Prompt examples:

• “Build my idea, then create 2 versions that solve this problem better”

• “Build me a working prototype of this flow”

• “Make a quick internal tool that does X”

• “Turn this PRD into something I can click through”

Tier 5: Shipping to prod

Yes, you, non-technical person. You can actually create code that goes into the core offering of your company’s product. Instant superpower and job security unlock.

Prompt examples:

• “Fix this bug”

• “Change this text”

• “Make the following changes to this flow”

Tier 6: Agents

This is truly the next level… but proceed with caution. Agents are great when you have strong context and lots of repetitive tasks. But they often require more setup, oversight, and cost than just doing the thing yourself. So it’s worth checking these out, but beware the hype. (And no, that AI chatbot on your site is NOT an agent.)

Prompt examples:

• “Every Monday, summarize last week’s sales and post to #revenue”

• “Each morning, pull yesterday’s signups from the DB and email me a digest”

• “Watch competitor pricing pages and ping me on any changes”

Throughout all of this, the most important skill you can have is creating clarity in all the chaos. If you’re feeling overwhelmed by everything that’s happening, remember that everyone else is, too. So don’t get sucked into the AI performance theater - if you don’t know what matters, what good looks like, or where to focus… AI just helps you do the wrong thing faster.

And if you’re looking for even more practical list of things to do tomorrow morning, I loved Claire Vo’s post about this:

So, you’re gonna lose your job. What are you going to do about it?

I hope you’ll take some time today to think seriously about what your role will look like in a year.

Even if you’re at the same company with the same title, I’d be shocked if you’re doing the same work.

After all, even I got ‘fired’ from my role as a manager a few months ago, to transition into doing IC work. You can’t avoid this kind of change. But there’s still time to prepare yourself. Ask yourself the tough questions, to decide what you really want to do and how you can prepare. Start your solopreneurship journey in some way, to give yourself career optionality. And start building an AI-native skillset, so you can thrive, no matter what.

Take one step toward each one of those. Then, go look at yourself in the mirror and tell yourself: ‘You’re gonna lose your job in 2027 and it’s gonna be your best year, yet.’

Edited by Jonathan Yagel (Fun fact: He built his new personal site with Lovable in an hour!)

But not just a regular, entry-level IC. I’m talking about employees with no direct reports who can do work that used to take a whole team (all while still getting paid like a leader). I’ve been calling this new role the High-Impact Individual Contributor.

And I’m not talking about this as a conceptual, hypothetical idea. I know this type of role exists because… I recently became one of them.

Sponsored by Glean: If you’re evaluating enterprise AI, this Model Context Benchmark report is worth a look. It breaks down why high-quality context matters, how Glean evaluated correctness and completeness, and why better context leads to more reliable AI results (Spoiler alert - Glean’s results were preferred almost 2× as often as ChatGPT’s and 1.6× as often as Claude’s!)

Sponsored by Amplitude: Build better products with our AI analytics platform. With powerful AI Agents embedded across our platform, teams can analyze, test, and optimize user experiences faster than ever.

The old promotion ladder was dumb

I still remember to this day, when I was at SurveyMonkey, a recruiter from Netflix reached out. I was super excited! The first question on the first call? ‘How many people are you managing?’ I only had 3 at the time, and this was for a team of 15, so I was immediately disqualified. That was the #1 metric. Not output. Not budget managed. Not ability to solve the key problems. Just headcount. WHUT?

So getting more people on my team became a hard qualifier on how much impact I could have.

This whole structure is so dumb. Just think about it:

We take people who are truly good at their craft, and we tell them to do a job that has nothing to do with that craft anymore. Our whole system promotes people out of the job they’re good at:

• Average craft, great people skills: could be a great manager, but don’t get promoted.

• Great craft, bad people skills: gets promoted to management anyway.

(Or, you incentivize politics by only promoting people who maneuver themselves… which is a disaster I’ve seen waaay too much and makes my blood boil, so we won’t talk about that here.)

But wait, there is more! These new managers are discouraged from maintaining their craft. Getting ‘in the weeds’ on projects gets described as micromanaging and ‘low value work,’ even though that’s the thing that was valuable enough for me to get promoted!?

When I hit the senior manager level, my boss literally gave me an ultimatum: ‘You have to stop. You can’t be spending all of your time on that stuff. I’m going to watch your Jira history. No SQL queries, no tickets! If you can do that for 6 months, I’ll know you’re ready for a Director position.’

So, I did it. Which… was actually amazing, at the time. Back then, managing people was the way to have impact, so that pushed me forward and taught me to achieve things through the people I managed. I’m super grateful to that boss for forcing me to step away from the keyboard.

But that’s not how the world works anymore.

What’s a HI-C (High-impact IC)?

A HI-C is: An individual contributor who can complete a project that delivers business value, end-to-end, on their own. Usually an ex-leader (manager, director, VP).

In contrast, traditional IC work looks like owning a piece of the puzzle. You are a key part of the machine, but you generally receive projects and hand off projects to someone else, without really being aware of what’s happening on the other side of those walls.

When I say ‘impact’ I mean: Directly affects the key metrics for the business. These all boil down to increased revenue or reduced costs. This is the full picture for the business, and the reason that senior-level people can have an impact (and get paid more) is because they see the whole thing. They’re responsible for getting the project from Point A to Point B that delivers measurable impact.

A director is usually managing managers who each have to get things from Point A to Point B, to get from Point A to Point C. And VPs manage directors, to get things from Point A to Point F. At the very top is the C-suite, who’s trying to get the whole company to move from beginning to end. The bigger your scope, the more money you make for the company.

The problem has always been that leaders were just connecting the dots and coordinating the teams. They weren’t actually doing the work. The whole point of management is to reduce the coordination costs and filter information up and down.

What if… instead, it looked like this?

That’s a Hi-C.

Yeah, should maybe be HI-IC, but that’s hard to say. And my millennial affinity for this juice brand is influencing my choice.

I’m living this world: I’m an IC again.

I’m telling you all of this because I’m living it, right now at Lovable.

This switch actually happened for me back in December. At first, I was building out a growth team. But then we decided to further flatten the org and decentralize growth, which meant... I moved back into an IC role.

I sit within the Product org and continue driving growth efforts… just by myself. We’re now hiring Growth PMs and engineers again, but they’ll sit parallel to me rather than reporting into me. My role is basically to parachute in and out of different projects and teams wherever I can create the most leverage. But mostly, I just have a huge block of IC work that I’m doing on my own. I’m running campaigns. Driving partnerships (did you see my latest LinkedIn x Lovable perk?). I’m coding my own growth optimizations and product updates. These are things that I used to rely on a large Growth org to do… now, I can do it all with just me, myself, and I.

For example, I recently built a prototype of our enterprise pricing page myself and shipped it to prod. In a previous org, that would have required a PM, a designer, and engineers, plus multiple rounds of iteration and at least a week of calendar time. More importantly, it would have pulled a team into something that might not even have been the right direction.

I will say: It messed with my head at first. I kept asking myself: Did I just level up, or did I just give up status? It’s so hard-wired into all of us: Must… have… team… to have… impact! Must have… important… title!

But having been in it for 3-4 months, I can say for sure: I’m lovin’ it. I have almost no meetings. I still do some coaching, but in general, I get to focus my work where I want every day. I get to attack the problems that I think are the most important and see how far I can take it, and just iterate with my own hands. I’ve always loved what I do, but a big part of the job was draining me (building decks for cross-functional alignment). Now, ~90% of my time is spent on the parts I actually enjoy - building.

Maximum IC impact, thanks to ‘Average Intelligence’

The High-impact IC role exists when a single person can complete the whole project from start to finish.

This is kind of like the senior-level ICs of the past (Staff Engineer, Principal Designer, etc)... but even more so, because those jobs were mostly about going super deep into one specific area, rather than being fully autonomous. I’d expect someone of that caliber to be the absolute expert on the code or design of a particular element of business… but not have them run a business function, end-to-end. Plus: That type of super IC role never really translated over to the Growth or marketing side of the house.

So what unlocks this now is (obviously) AI.

Here’s the thing: I think of AI as ‘average intelligence’ (heard this from Ravi Mehta) - which sounds like an insult, but I don’t mean it that way. It’s a really big deal to be automatically at the level of an average designer, an average marketer, an average product manager. Again: It’s not great or innovative or actually differentiated. But you don’t need everything to be at that level, right away.

A lot of times, ‘average’ plus your own craft is all it takes to get a real thing out into the world.

So now we have no excuse to be really bad at stuff - we can all be at least average very easily. Which means we can get things off the ground without needing a full team upfront. You can build, pressure-test, and validate ideas by yourself, and only pull in real experts once you’ve got some momentum.

And when you can ship faster, you can learn a lot faster, too. When building was slow and expensive, there was way more pressure to get everything ‘right’ upfront. That’s why everything traveled up and down the chain. IC → manager → director → VP → back down again. The job of a manager was mostly to review, approve, and de-risk decisions before anything got built.

But when you can build and test something in hours… It’s now becoming cheaper to just try something than to debate it. The cost of the mistakes is much lower now. Which means you don’t need as many layers of decision-making. You don’t need as much escalation. You don’t need someone whose primary job is to sit in the middle and approve or redirect.

Escaping the Coordination Loop

The existing system is all built around coordination, which incentivizes management. Which pushes people into management positions. As the company grows, the rows keep stacking up and up until most of the organization is just coordinating other people’s coordination. I mean... middle management, anyone?

The good news is that the Hi-C structure can do the opposite, and create a positive loop:

When an individual can run a project end-to-end, they don’t need as much coordination. This means you don’t need as many managers. Which means that your comp budget naturally flows to the people having the biggest impact, which are now ICs. Which means that you can keep your best people in the roles that they’re really good at.

How do companies make this happen?

A single person that can have department-level impact? Yes, please. Every company wants that. But… how do you get it? Two things:

(1) The biggest blocker is ungating access to information. This kind of system only works if this type of IC has the same level of context that a senior leader could get. If they’re only getting the bits and pieces that get passed down the line by a thousand gatekeeping middle managers, it’s not gonna work.

To get this really incredible level of impact out of people, you have to trust them. And most large orgs just don’t trust their employees. So, this kind of thing is working at the small AI-native companies, not just because of the natural preference for AI tech, but because the bureaucracy hasn’t emerged around a group of middle managers who need to defend their own value through constant (slow) coordination.

If you want to see if your org can handle this, invite some of your high-potential junior ICs to step up to lead full-stack projects. Wherever they get stuck is going to be a bottleneck for this kind of role.

(2) Recruit people in leadership positions to join you as an IC. Seems obvious, but this almost never happens. What we’ve noticed in pitching this at Lovable is that, yeah, some say ‘Absolutely not. I need another VP title.’ But some of them say, ‘Hells yeah, this is amazing!’ This has never been a conversation because in the past, IC fundamentally meant a reduction in impact… and a massive reduction in salary. But it doesn’t mean those things anymore.

No more mandatory Managers

I hope literally every person reading this is breathing a sigh of relief.

For managers who wake up excited about coaching, team development, cross-group coordination… stick with it! If you love the work, there will still be chances to support these turbo-powered Hi-C folks.

But for the rest of us: Can we retrain ourselves to think of the management path as truly optional?

Traditional ICs: You don’t need a team under you to have a big impact on your company (or your net worth). Keep developing your craft, and figure out how to connect what you’re doing to your organization’s core metrics.

• Marketer? Don’t just plan the campaign. See what customers have been asking for from the product. Figure out if you should be pushing for new users vs. existing users. Build the actual in-product promo experience. Make a list of improvements to the features, and code them yourself.

• Product? Don’t just have your marketing counterparts handle everything related to distribution. Learn about how the market and customers talk about the product. Understand the different demand drivers from different segments (enterprise vs. SMB). Figure out the most effective marketing channels for talking about what you’re building.

Leaders: Stop juggling agendas and decks and quarterly plans, and get back in touch with the craft that started your career. Look at a valuable project on your team’s roadmap that might not happen this month. Instead of delegating it, try to do it yourself (deploy AI as needed). Once you see how far you can take something, you’ll start to have the ah-ha moment that I did. And maybe you can fall back in love with your work again.

But now that serious impact (and the comp that comes with it) can be available without the headcount… let’s all agree to go spend that energy building something we love instead of climbing something we hate.

Edited by Jonathan Yagel.

Most companies seriously underinvest in monetization as a growth lever.

(Side note - when did Sessions get so big??? 6K+ people showed up and I’m hearing it’ll be even bigger next year. Wow. Definitely not an event to miss.)

They obsess over acquisition channels, activation funnels, retention curves, and roadmap prioritization, while pricing and packaging get treated like a dangerous object in the corner of the room. Something to avoid touching unless absolutely necessary.

That mindset is costly.

Because monetization is not just how you capture value. It is one of the most powerful ways to create growth, improve customer experience, and adapt to changing markets.

At Lovable, one of the biggest lessons from the last year has been that pricing should be treated much more like product than policy.

Read more

Hi everyone - Most of you are Growth or product people, which means that one of your closest collaborators is probably: Product Marketing. As I’m seeing roles start to merge and overlap, this is an area you should definitely be aware of. What are AI-native PMMs doing and how can you collaborate best with them? What skills do they have that you should maybe learn? I asked our incredible Core PMM from Lovable, Gemma Casey-Swift, to write this guest post to fill you in. Enjoy! - Elena

How product marketing works these days is unrecognizable from even a few months ago.

Some of it’s the available AI tools, some of it’s the speed of feature releases, some of it’s the shifting job roles. But if you’re trying to launch products in late-April 2026 (it may be different next week!), please reconsider what you think you know about PMM.

One of the biggest changes? Product Marketers used to be a bottleneck. Building a lot of elaborate processes (yes, I said the p-word, I’m sorry) that seemed like they were slowing things down.

Now? Well, now the PMM team is a crucial part of helping your org go fast. Here are a few key ways that product marketing at Lovable has built foundations that the rest of our team needs.

(If you’re a Growth or Product person… hey, friends! Send this to your PMM to make sure they’re tracking. But also maybe read it yourself? Spoiler alert: The more you can do your own product marketing, the faster your career and company will grow.)

Let’s do this!

Elena’s Picks

Hey, I’m trying out a new format for my sponsors! I’ll be specifically recommending products and services I personally use and love.

→ Stripe: Ideal for payment processing and all-things monetization.
I’m giving a talk at their big conference, Stripe Sessions… tomorrow! I’ll be sharing how Lovable turned monetization into a growth lever. Come check it out!
🎁 Come hang out: Register for Stripe Sessions.

→ Customer.io: Ideal for highly-targeted email campaigns based on user behavior.
We use them at Lovable and I’m a huge fan. It’s so flexible and intuitive to use that, before we had our first email marketer, anyone on our team could launch an email campaign! And it was actually cohesive. Now that we do have an email marketer, they’re doing amazing things with it. A product that can serve varying degrees of experience like that is so hard - but Customer.io has nailed it.
👀 Check it out: Learn more.

→ Lovable: What can I say? Yes, I work here. But… I work here because I was obsessed with the product, first. It’s the best way to build an app or site, even if you have no technical experience. Try it out!
🎁 Offer: Get a year of Lovable for free! (For paid subscribers to Growth Scoop)

1 - Enable your team to launch stuff, themselves (vs. expecting everything to go through you)

This is probably the biggest switch. The reality is that we are launching updates wayyy too fast for a traditional PMM-driven launch cadence. In old school companies, you’d have a big Tier 1 launch at your user conference… once a year? Maybe twice if you’re aggressive.

Now, we launch updates that could be tier 1 almost every other week. So, how do we keep up this pace? Definitely not by making everything come through PMM every time. If we did that, we’d always be the bottleneck.

So, instead, we built out the infrastructure to help our team launch stuff on their own.

We tiered launches into a few main buckets:

• Big launches (still Tier 1): These are still big releases (or round-ups of multiple releases) with big moments and active coordination across the teams. Marketing still runs these.

• Tier 2: Still run by product marketing alongside a PM but generally more fluid with a more flexible bill of materials. We often amplify these further post-announcement if the launch is getting good traction.

• Everything else: If it’s not specifically designated for a tier 1 or tier 2 launch, the people building it launch it themselves. Our job as the product marketers is to give everyone the resources they need to do this.

This is the key to going fast. The whole point is to create a culture where people are encouraged to launch their own stuff without PMM being precious. You lose a little messaging control, but you gain momentum. And honestly, some of the content our eng and product team shares goes more viral than any perfectly curated post ever could.

So, what kind of infrastructure are we actually building?

2 - Define your ICP based on actual data (vs. based on brainstorming or wishful thinking)

Providing a clear, consistent Ideal Customer Profile (ICP) is one of the most valuable things a PMM team can provide to their organization. Unfortunately, most of these are all over the place and based on some hand-wavy hypotheticals.

Now, no one has time for an ICP that doesn’t drive value.

How do you get that? The good news is that this profile actually already exists. It’s just buried in your CRM, usage data, support tickets, and that spreadsheet someone in sales has been carefully maintaining.

Here’s how we pull the data with help from our awesome insights team:

• Product data: We dug into usage, retention, ARR, and conversion across the full user base.

• Builder surveys: Captured roles, needs, decision-making, and “definition of done.” Paired responses tracking expectations vs. outcomes across a build cycle.

• Deep-dive interviews: Tapped further into what people were building, why, their emotional state, workflows, needs, struggles, and dreams.

• Internal AI agents (built on Lovable, live in Slack): Built by our insights team and connected to our full library of user calls, surveys, and interviews. We ask questions in natural language and get back themes, quotes, and patterns on demand.

The biggest shift for me was realizing our ICP is genuinely “everyone,” and you can’t ignore that.

My old instinct was to narrow: pick one persona, one segment, one use case, and aim everything at them. But when the product is category-defining and/or general-purpose, demand can show up from solo-builders to huge enterprises all at once. Even if this makes product marketing and prioritization hard, you can’t ignore clear signals.

This is also why ICP reviews can’t be an annual ritual anymore. PMMs traditionally set an ICP in January, build a deck, and let it calcify. That cadence won’t work today. We review which segments are growing, converting, expanding, and churning weekly. Most quarters there’s a small adjustment. Some quarters there’s a real one, and catching it early changes what marketing, product, and sales should be doing right now.

Naming the ICP is one thing, the actual job is making the rest of the company move with it. PMM is constantly refreshing core messaging, channel mixes, targeting and talk tracks.

We package ICP insights into clear updates and make the rounds on Slack channels, all-hands, team meetings, 1:1s with leads, etc.

3 - Constantly interview customers (not just to start, or when you’re forced to)

Most product marketing teams either rely on other teams to gather direct customer feedback… or they just don’t get it at all. But if you’re going to build fast and launch fast, you need to be talking to your customers a LOT. Do not let sales or CS block you (this was the bane of my life for years).

When I joined, me and our Growth PMM immediately linked arms and packed our calendars with back-to-back interviews. We went to every team and asked them a million questions about our users.

On each customer call, we asked the same five questions every time:

1. “What were you trying to do when you found us?”

2. “What did you try before Lovable?”

3. “Walk me through the moment you decided to sign up.”

4. “What almost stopped you from using it?”

5. “How do you describe Lovable to other people?”

We recorded everything (with permission) and looked at trends and patterns.

This kind of information is the lifeblood for any product-based business. Having real-time feedback on what’s working and what’s not helps everyone go faster. So, what were the pain points that users mentioned?

• “Hiring developers is expensive and slow”

• “I tried no-code tools but hit limits immediately”

• “I’m not technical but I have ideas”

And it wasn’t just the themes. A big part was making note of the exact language they used:

• “It gives me superpowers” (not “it writes my code”)

• “I had a working prototype by lunch” (outcome-focused, time-specific)

• “Finally something that just works” (cognitive ease)

And the moments they got excited about:

• Watching their ideas come to life before their eyes for the first time

• When they shared what they built and someone said “You built this?!”

• When they realized they didn’t need to hire anyone

• When something they thought would take weeks took 20 minutes

And the questions that kept coming up:

• “Is this actually production-ready?”

• “How do I work with my brand?”

• “Can I export the code?”

Back in the day, these insights would probably just sit in a doc. Now, they’re powering tools that every team at Lovable can use, including:

• An interactive messaging house (built in Lovable, naturally). We took the core messaging and turned it into a living document inside a Lovable app. Not a deck. Not a Notion page that goes stale in a week. An interactive messaging house anyone in the company can navigate, search, and pull from by audience, by archetype, or by use case.

• An AI research assistant. Remember the agent I mentioned earlier? Our data scientists built it so anyone at Lovable can just ask things like “How do solo-founders describe the product?” or “What’s the #1 frustration for small teams shipping to production?” and get insights from every interview, survey, and call we have on file.

This means every team can get more accurate, higher-quality context when trying to communicate with current or potential customers. For me, when I was writing landing page copy a few days later, I literally pulled exact phrases:

• Landing page headline: “Ship your ideas in days, not months” (exact phrase from user interview)

• Social proof: “I’m not technical but I have ideas—Lovable is perfect for people like me” (real quote)

• Campaign copy: Added “Your co-founder doesn’t just build, it runs the business with you” (based on how people were describing the product)

4 - Create a simplified messaging framework (vs. making everyone invent messaging from scratch OR read an exhaustive 25-page treatise)

Does your company have a central source for messaging? When I joined, I asked about this… and it turns out one existed from a few months prior but no one remembered what it said.

You’ve probably seen it before: The PMM team beats everyone over the head with messaging documents that are constantly updated and also dozens of pages long, full of details that seem oddly specific and still somehow hard to understand.

But it never hurts to start over, especially if you’re new. I asked everyone on the team: “Show me how you describe what we do.” And that led me to:

• Pitch decks (multiple versions)

• A bunch of different Google docs

• Landing page copy (homepage, pricing, feature pages)

• Email templates

• Links to tons of social content and videos

• Customer-facing help docs

• Interview recordings

• A bunch of deep Slack threads

And we started looking at the patterns:

• People didn’t see Lovable as a coding tool - it was becoming a “full co-founder”

• Everyone mentioned speed (“fast,” “10x,” “minutes not months”)

• Split on whether to lead with AI or outcomes

• Technical teams said “full-stack apps,” non-technical teams said “just describe what you want”

Which also forced us to consider some unresolved conflicts:

• How technical are our audience really?

• What do they really want?

• Is the magic the AI or the speed?

• Are we selling software or selling outcomes?

After some hard conversations, I was able to build a single messaging hierarchy with our core positioning, messaging variants for technical and non-technical users, use cases, and proof points. And again, the thing that helps everyone go fast with this kind of thing is… actually getting access to it. So I presented it to key stakeholders, got feedback, adjusted, and circulated.

Within two weeks, people stopped asking “How should we describe what we do?” The marketing team updated copy across all of our channels. The product team echoes the same messaging in roadmap planning sessions. The founders reference it in company updates. Everyone’s able to move quickly, because they’re not having to build messaging from scratch every time.

BONUS: PUT IT WHERE THEY CAN ASK IT

One of the biggest time sucks for any organization is trying to find the right information. It’s all probably out there… it’s just scattered across Slack threads and Google Docs no one can find, or the PPT decks with “FINAL - Final Actual v2.8” in their titles. If you don’t know whether something is current or accurate, you can’t trust it. And none of the work PMM does matters if it’s buried.

The old version of fixing this was “put it where people can find it”. By that I mean creating and distributing single source of truth documents with links to relevant resources so the same question doesn’t get asked twenty times. That’s still table stakes, and it’s still one of the highest-leverage things a lot of PMMs aren’t doing.

But the bar is moving fast. The next version is “Put it where they can ask it.” Instead of hunting for the right doc, anyone in the org should be able to ask a question to an agent or LLM chatbot, in natural language: “What’s the messaging for the payments launch?” “Who’s our ICP for enterprise pilots?” “What are the top 10 most used features right now?” The goal is to be able to ask these questions and get an answer on demand, wherever they work.

We’re already running a network of specialized agents on Lovable, which are basically small company brains built to help people with specific areas of their work when they need it. Next up: a launch agent connected to our calendar, messaging briefs, and asset libraries, so anyone can ask about upcoming launches, find the information they need, and talk about it.

Conclusion: Product Marketing at ludicrous speed

So, who has time for product marketing? Turns out everyone does when product marketing starts thinking more like an infrastructure, not as a gatekeeper.

Providing a clear ICP grounded in real data, easy access to that live data, self-serve launches, and information that is findable, askable, and usable by anyone in the org. Done right, PMM stops being the team you wait on and starts being the team that makes everyone else faster.

If you’re a PMM reading this, throw out the parts of the playbook that were really about controlling work rather than enabling it. Build the foundations, then get out of the way.

If you’re a Growth or Product person reading this, your closest collaborator just became one of the highest-leverage people in your org, so don’t outsource product marketing thinking, borrow it. Sit in on customer calls, pull from the messaging doc, ask the research bot. The more you can do your own product marketing, the faster you (and your company) move.

PMM isn’t the bottleneck anymore. In an AI-native company, it’s one of the engines. so go find yours, or build the one your team’s been waiting for.

Want more from Gemma? Check out her awesome new substack:

Edited by Jonathan Yagel.

It’s the thing that makes the whole company go around - marketing and sales build their entire personality around it, founders try to change it every week, and engineers nod while secretly thinking ‘this feels made up.’

Define your customer → build for them → grow. Easy. In theory.

But there’s an assumption baked into all of this: Your user is human. I think that assumption is breaking.

As agents begin to interact with products on our behalf - often via protocols like Model Context Protocol (MCP) - your ‘user’ may never actually touch your product.

Which changes pretty much everything.

In the past, even API-first products had a clear answer. Your ICP was still a human, usually a developer. Someone chose your product, integrated it, and used it. But now, LLMs will definitely be involved in every step of that process, and may even replace it. This is the question I think every company should be asking:

Wh…

Read more

We went from on-prem to cloud. Then came digital transformation, where every company suddenly needed to become a software company (whether they had any business doing so or not). Entire stacks got rebuilt - .NET stopped feeling inevitable, Python spread everywhere, and everyone got weirdly excited about Next.js. Mobile came in hot (B2B, naturally, is still deciding whether a phone is a real device). Recurring revenue became our SaaS religion. Remote work exploded, then partially un-exploded. Product-led growth rewired how companies scaled.

Each wave felt big. AI feels 10x bigger. And 10x faster.

That speed is the part I can’t quite process. There are days where it genuinely feels like I need a full-time job just to keep up with what’s happening. And even then, I’m not convinced I am. Am I just getting better at being slightly less behind?? Haha.

And that’s really the feeling underneath all of this: behind.

Not ‘I should probably test that new tool’ behind. More like ‘am I already operating on an outdated mental model?’ behind.

There’s also a weird social dynamic to this whole thing. Everyone has a system, a stack, a workflow that supposedly changed their life, saved them from the burnout, and probably whitened their teeth. It creates this illusion that everyone else already knows what’s going on.

So you hesitate to ask basic questions, because it feels like you’re the only dumb one who doesn’t get it.

Which is almost certainly not true. But it feels true. And the result is a lot of fake confidence theater, a lot of noise, and very little shared truth about what’s actually working.

Then there’s the part that’s harder to admit: a lot of what I spent the last decade learning is losing leverage.

Growth, marketing, product management, sales - these used to feel like crafts. You built intuition over years. You learned what great looked like. You got good at pattern recognition. You earned judgment and respect by grinding through it all.

AI is flattening a lot of that.

It’s a weird experience to spend ten years becoming excellent at something only to watch a 22-year-old produce a suspiciously solid version of it in 14 minutes. Not the best version. But enough to make you sit there like, wow, okay then. And you don’t really get to process that shift. There’s no mourning period. You just move on to the next thing, because everything is changing too fast. (and don’t even get me started on how non-tech people are supposed to keep up with any of this - ‘mass adoption’ feels almost comical.)

Which forces a much more uncomfortable question: if your identity was tied to being good at X, what happens when X gets automated?

Now, senior people still have some advantages, but some of the old signals of seniority are getting weaker fast. The hierarchy flattens. The ladder gets weird. Experience still matters, but not in the clean, linear way we got used to.

Honestly, I think that a lot of that is good. A lot of old hierarchies deserved to die. RIP politics and middle-management. Still, it’s disorienting as we are going through it.

And of course, the reward for becoming more efficient is exactly what it has always been: more work. If AI makes you 10x faster, nobody says, ‘Amazing. Please enjoy your afternoon!’ They ask why you’re not doing 10x more. Productivity gains get absorbed immediately by the system, which is very on-brand for tech.

And underneath all of this is the part we really don’t want to talk about: the economics.

If output goes up while the cost of producing that output goes down, what happens to compensation? What happens to leverage? What happens to the premium attached to knowledge work when more of that work becomes abundant? What happens to company structures when fewer people can produce dramatically more?

My guess is the value shifts.

Away from execution. Away from being the person who can grind through the work manually. Toward taste, judgment, prioritization, and orchestration. Toward deciding what’s worth building, not just how to build it.

Which sounds elegant until you realize those are harder skills to build patterns on and really really hard to teach/learn.

The last question I keep coming back to is whether we are the last generation that built our careers around software as a medium.

We leaned all the way in. We learned the tools, the systems, the playbooks. We rode cloud. We drove digital transformation. We built careers on knowing how to operate software better than other people. We are also, unfortunately, the sandwich generation that has to explain software to our parents and now AI to our kids.

But if software starts getting built, configured, and operated by AI, what happens to the people whose edge was being good at creating software?

I don’t know. And I’m not convinced anyone else does either, despite the amount of confidence cosplay happening online.

And maybe I’m overblowing it. Or maybe this is just what it feels like to live through a real industry shift while still being expected to perform like nothing has changed.

The only thing I feel confident about is that the advantage is shifting.

It’s shifting toward the people who can see clearly, choose well, and adapt faster than the ground is moving beneath them.

No pressure everyone. Everything will be okay. (RIGHT?)

Usage.
Conversion.
Retention.
Revenue.

A much less comfortable question: do users actually give a shit?

Because a lot of “healthy” products are really just products people tolerate:
their boss told them to use it, they have not found a better alternative yet, or switching feels too annoying.

None of that means they love it.

And the second a real alternative shows up, they’re gone. There goes your retention.

If you want durable growth, the distinction between “I have to use it” and “I love using it” matters a lot.

Read more

Here are a few that we’ve recently tested… which all yielded a meaningful lift!

Plus, 2 laws of retention improvement that everyone keeps forgetting.

Read more

The kind where you squeeze customers a little harder every quarter, slowly wreck trust, burn the brand, and then six months later everyone’s in a meeting asking, “Why is retention and growth suddenly bad?”

And here’s the hard truth: in this age of technological trAInsformation, that kind of revenue chasing is straight up fatal.

I’ll go even further: If revenue is your North Star, your company is destined to die.

Read more

And sure, we can debate what ‘viable’ is supposed to mean, but let’s be real - MVP devolved into “minimal functioning product” because ‘viable’ got watered down to ‘it performs bare minimum.’

And that’s because dev work was expensive. Like very expensive. But we live in the new reality now where dev costs are collapsing. Everyone’s a builder now. Software functionality is becoming a… commodity. So an MVP just means you built another tool people tolerate until they re-create it themselves or a better one comes along.

An MLP is different: earliest version that’s genuinely lovable. The kind of product where people feel it. It’s fast, obvious, opinionated, and gives you that ‘wait, this is actually… nice’ moment.

But most people still don’t get why this matters so much right now.

When I bring up MLP, I always get pushback. ‘You’re describi…

Read more

I agree with the core premise of his article. Disruption is coming, and it’s going to be massive. But I think most people are framing the conversation the wrong way. Everyone is stuck debating whether AI will take their job. The answer is almost certainly yes (I know, I know - it’s terrifying). That part isn’t interesting anymore.

The more important question is this: What if this is also the biggest opportunity of your career?

Right now, there’s a short window where individuals can get absurdly far ahead by becoming AI-native. Not “AI-aware.” Not “occasionally uses Claude.” AI-native. And that window is measured in months, not years.

Who should actually be scared?

It’s a terrible time to be a company or an exec. But it’s a great time to be an individual contributor.

If you’re a founder/ceo or exec running a bloated, slow-moving org that spent decades layering process, coordination, and middle management to handle complexity AI is about to flatten, you should be sweating. Companies that need 26 meetings, four approval chains, and a Gantt chart from 2019 to ship a button color change aren’t careful. They’re exposed.

At this point, many companies are only alive because of distribution, network effect, or data lock-in. The product itself is no longer the moat.

But if you’re an individual stuck in a role you don’t love, underpaid, boxed in, spending your days in meetings that accomplish nothing, writing docs nobody reads, and waiting for your ideas to get prioritized on someone else’s roadmap, the bad news for them might actually be the best news for you.

The demand for AI-native employees is exploding. Not in some abstract, future way. Right now. Companies are starting to care less about fancy titles, perfectly laddered careers, and padded resumes, and more about who can actually get things done in this new environment. People with good taste, real business sense, and, most importantly, high drive. In the next year or two, I’m convinced being AI-native plus those qualities will matter more than years of experience, brand-name logos, or titles that only made sense in a slower world.

Let me make this concrete

At Lovable, we have a Slack channel called #feedback where anyone can post things that feel broken or annoying in the product. Normally, devs would jump in and fix them.

But lately, I started noticing something different.

Instead of devs writing the fixes, I kept seeing messages like, “Hey Cursor, fix this” or “Claude Code, can you handle this?” And then… the fix would just happen. A PR would appear. A quick human review later, it was live.

So next time I was ready to post in the channel (CTA button had the wrong color), I had a wild thought: could I fix this myself?

I don’t have an engineering background. I didn’t even have a GitHub account. But I remembered a teammate (Benjamin) I’d seen do this, so I asked him to show me how it worked. He helped me set up GitHub, Claude Code, and Cursor.

I took a screenshot of the button, told Cursor what was wrong, and asked it to fix it.

It immediately found the issue, fixed it, and opened a PR. The engineer reviewed it and merged it.

I had just “written” code to production. In minutes. Something that would normally take tickets, meetings, and weeks of waiting… I just did it. Not because I became an engineer, but because the gap between intent and execution collapsed.

Honestly, it felt like the biggest unlock for me since discovering Lovable. Equal parts excitement, an undeserved sense of superiority, and just enough danger to be intoxicating.

For individuals, this is leverage. And it’s happening right now.

This post is brought to you by Salespeak’s free new tool, IsYourWebsiteReady.ai:

AI agents influence buyer decisions without triggering analytics sessions.
ChatGPT. Claude. Perplexity. Already accessing your site.

Your analytics won’t show it. Agent Analytics does. Free.

👉 Reveal your AI activity

Get your head out of the sand

A lot of people’s response to AI right now is critique. Pointing out where it fails. Where it’s wrong. Why it’s overhyped. Screening for em-dashes. Every new model drop turns into a race to prove what it can’t do.

I get it. If something threatens your role, tearing it down feels like self-defense.

But that fear only makes sense if the current setup is worth protecting. And watching this play out, it feels like people are desperately defending a system that was already failing them. Meetings instead of making. Process instead of progress. “Work about work” that creates no value.

AI is not just automating tasks. It’s removing permission. You no longer need to wait for headcount, budget, or alignment to create something real. That’s why companies are nervous. A motivated individual with curiosity and initiative can now do what used to require a team.

This is why there’s a window right now.

Early adopters are compounding advantage. They’re learning faster, shipping faster, and building confidence faster. Once AI-native workflows become the default, that edge disappears. Right now, curiosity beats credentials. But that won’t last when AI skill becomes commoditized.

I’ve spent years writing about this stuff. About how full-time employment stability is a myth - were you 100% safe from layoffs, even before AI? About how tying your identity to your job title is a trap. How the leadership ladder has a dark side that pulls people away from work they actually love and into meetings, politics, and people management they hate. How I quit full-time roles to build career optionality - how owning your skills, your reputation, and your ability to create value independently is the real security. And how I literally want to eliminate my own job, because so much of Growth has been stagnating and I think AI could free us up to do better, more creative work.

Things you should do this week

First, lock in and try to vibe-code a rough version of your own company’s product. Block a few hours (using Lovable, obvi, LOL). No perfection. Just see how far you get.

Don’t start with edge cases or scale. Start with the core thing your product actually does. Most companies have two or three workflows that matter and a long tail of noise wrapped around them. Focus on those. See if your users can create their own personal tools solving for that workflow.

If AI gets you uncomfortably far in a single day, that’s a signal for you of what’s to come.

Second, make AI your default for everything. AI should be your starting point for everything: writing, thinking, designing, coding, planning.

The culture around this is backwards. We spend so much time shaming people for not deleting them em dashes... Oh how wrong we are here. The real question now is why someone didn’t use AI in the first place.

Manually doing work that AI can do better, faster, and cheaper isn’t craftsmanship. It’s wasted time. And in a world where leverage compounds weekly, wasting time is the most expensive mistake you can make.

Third, if AI couldn’t do something yesterday, try it again next week. AI improves in weeks, not years. Once AI began writing its own code to improve itself, the pace of capability growth went exponential. “We tried that and it didn’t work” is the fastest way to fall behind.

Fourth, start a side project. Not because you’re quitting tomorrow, but because building is the fastest way to internalize what’s changing. The tools are cheap or free. The barrier to entry is gone. Maybe you can even make money from it.

Fifth, ask the AI-native person around you for help. They’re waiting to show you. This stuff is contagious once you see it. Don’t do this:

Stop reading. Start shipping.

No one wants to lose their job. That fear is real. People depend on your income. I’m not dismissing that. But ask yourself honestly: do you love your job so much that you want it frozen exactly as it is? Do you want your kids to have it?

You need a job. But do you need this one, in this form, forever?

What if this moment isn’t about falling behind, but about finally getting unstuck?

The biggest risk right now isn’t AI taking your job. It’s watching other people pull ahead while you wait to feel ready.

So stop scrolling. Open Claude, ChatGPT, Cursor, or Lovable. Try something. It doesn’t have to be perfect. It doesn’t have to be big.

Just start.

Edited by Jonathan Yagel.