The Smartest People in the Room Are No Longer in the Room

In partnership with

AI Agents Are Reading Your Docs. Are You Ready?

Last month, 48% of visitors to documentation sites across Mintlify were AI agents, not humans.

Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.

This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.

Your docs aren't just helping users anymore. They're your product's first interview with the machines deciding whether to recommend you.

That means: clear schema markup so agents can parse your content, real benchmarks instead of marketing fluff, open endpoints agents can actually test, and honest comparisons that emphasize strengths without hype.

Mintlify powers documentation for over 20,000 companies, reaching 100M+ people every year. We just raised a $45M Series B led by @a16z and @SalesforceVC to build the knowledge layer for the agent era.

Make Your Docs Agent-Ready

If you have spent the last six months using AI to write emails faster, you are the group that is going to lose.

The people who are winning right now are not the ones who work the hardest. They are not the most experienced. In some cases, they are not even the most technically skilled.

They are the ones who figured out how to think alongside AI before everyone else did.

And the distance between them and the rest of the field is growing every quarter.

What Spotify Showed Us Without Meaning To

When Spotify's co-CEO mentioned on an earnings call that his best engineers had not written a single line of code since December, most people heard a productivity story.

It was actually something more unsettling.

Those engineers built an internal system called Honk, powered by Claude Code, where an engineer can type a message on their phone during their morning commute and receive a completed, tested, production-ready feature before they sit down at their desk. Fifty new features shipped in 2025 through that workflow.

But here is what nobody talked about after that headline.

Those engineers did not get there by working harder. They got there by completely changing what they considered their job to be. They stopped thinking of themselves as people who write code. They started thinking of themselves as people who design systems that produce outcomes.

That is not a productivity upgrade. That is a professional identity shift.

And most people have not made it yet.

The Quiet Divide That Is Already Happening

In almost every knowledge-work field right now, two groups are emerging.

The first group uses AI as a faster version of what they were already doing. They write emails faster. They summarize documents faster. They produce first drafts faster. Their output increases by maybe 30 or 40 percent, and they feel good about that.

The second group uses AI to do things they could not have done at all before. They build systems that compound. They take on projects that would have required a team. They work at a level of ambition and output that makes the first group look like they are still using dial-up.

Both groups are using AI. The gap between them is not a tool gap. It is a thinking gap.

The first group asks AI to help them do their job. The second group asks AI to help them reinvent what their job even is.

Here is what that looks like in practice. Two copywriters started 2025 with identical résumés. The first used AI to write more ads per month and by December was producing four times the previous year's volume. The second used AI to build a system that generated 200 variants of every brief, tested them, and surfaced the three that performed. By December she was not selling ads. She was selling access to the system, on retainer, to three clients at once. One has a salary. The other has an asset. Same tools. Different question about what the tools were for.

Why Most Advice About AI Is Wrong

The conversation about AI in the workplace has settled into a comfortable and mostly useless frame.

Learn the tools. Master the prompts. Stay ahead of the curve.

All of that is true and none of it is the point.

The people who are genuinely winning with AI are not winning because they know more tools. They are winning because they have a different relationship with the question of what their time is actually for.

When AI can handle execution, the most valuable thing a person can do is figure out what is worth executing on. That sounds simple. It is actually very hard, because most professional training is about execution. Most performance metrics measure execution. Most workplaces still reward people for how much they produce rather than for how clearly they think about what should be produced.

The shift AI is forcing is not a technical one. It is a strategic one. And most organizations are not designed to support it.

The Leverage Problem

Here is the thing about leverage that does not get said often enough.

Leverage is not just about doing more. It is about who benefits when you do more.

An individual contributor who learns to use AI well can increase their personal output dramatically. That is real. But if they are working inside a system that captures that output without rewarding it, they have simply made someone else richer while adding very little to their own position.

The people who are actually building something with AI leverage are the ones who own the output. Founders. Freelancers with the right client relationships. People inside organizations who have negotiated to be compensated on results rather than time.

Everyone else is getting more productive on behalf of a system that was designed before AI existed and has not been redesigned since.

Picture the junior consultant who starts delivering in two days what her seniors used to deliver in two weeks. She assumes she will be promoted faster. She is not. Her quota goes up instead. Same salary, same title, three times the throughput, all of it absorbed by the firm. The productivity gain is real. The question is who keeps it.

That is not an argument against learning AI. It is an argument for thinking carefully about the structure of the situation you are in before you invest in becoming dramatically more efficient inside it.

What Actually Transfers

If the tools keep changing, and they will, the only thing that compounds reliably is judgment.

Judgment about what problems are worth solving. Judgment about when an AI output is good enough and when it is subtly wrong in ways that matter. Judgment about which parts of a workflow should be automated and which parts should stay human because the human element is load-bearing in ways that are not immediately obvious.

That kind of judgment does not come from using AI more. It comes from thinking seriously about what you are actually trying to accomplish and why the current approach to it might be inadequate.

The engineers at Spotify did not stumble into building Honk. They had a clear view of what they were trying to produce and worked backward from that to figure out what the AI needed to do. The tool came second. The clarity came first.

This has happened before with a smaller tool. In 2007 the iPhone put a serviceable camera in every pocket. The photographers who collapsed over the next five years were the ones whose value proposition had been “I take pictures.” The camera stopped being the asset. The ones who came out stronger sold direction, composition, judgment about which moment was worth capturing. The phone never threatened them, because the phone could not do the thing they were actually being paid for. The question worth asking yourself is whether your current professional self-description survives the same test.

The Cost of Sitting This Out

Look at the tech layoffs of 2025 and the first half of 2026. They were not announced as cost cuts. They were announced as reorganizations, restructurings, focus shifts. The pattern that is probably forming underneath, when you look at who stays and who goes, is the same one everywhere. The middle of the team leaves (the fiftieth percentile). The top stays (the ninetieth). The top now does the work of both, with AI absorbing the difference. Nobody is going to say this on a conference call. The headcount math will say it for them.

The honest version of where this goes, if you do nothing, is not that you lose your job tomorrow. It is that two years from now you are probably doing the work of three people for the salary of one, inside a structure that has quietly stopped promoting anyone in your position. That is not a dramatic ending. It is a small one. Which is exactly what makes it worth taking seriously now, while the move is still cheap.

The Question Worth Sitting With

The best use of this moment is not to become an expert in any particular AI tool.

It is to figure out what you would do if you had ten times more capacity than you currently have.

Because if you do not have an answer to that question, more capacity is not going to help you.

The people who are pulling ahead right now had that answer before the tools arrived.

If this raised a question you have been thinking about differently, send it back. The most interesting threads always come from readers who are sitting with something specific.

Thanks for reading. See you on Thursday.

10x the context. Half the time.

Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.

Try Wispr Flow free