For 15+ years, I was the knowledge worker. The brief would land, I'd disappear into research for a week, synthesise everything into something nobody else could see, and come back with the insight. That gap between the brief and the answer was where my value lived. It was where all of our value lived.
That gap is now approximately 180 seconds long.
The business model nobody talked about
The knowledge economy wasn't built on intelligence. It was built on asymmetry: knowing more than the person across the table. Strategists, consultants, analysts, lawyers, journalists, researchers -- every one of these roles was underwritten by the same assumption. Access to information is unequal, and the people with better access charge for it.
This worked for decades. It created genuine value and funded entire industries. The strategist who'd read every category report, spoken to every consumer, synthesised every signal. The lawyer who'd memorised the precedents. The journalist who'd cultivated the sources. The value wasn't always the raw information itself -- it was the time, the rigour, and the judgement applied to it.
AI hasn't just made the information cheaper. It's collapsed the time-to-knowing so completely that the gap has effectively closed. A startup founder in Lagos now has access to the same strategic thinking that used to cost £200k in agency fees. A first-generation student can research a legal question like a McKinsey associate. This is the biggest unlocking of capability in a generation. It is also, genuinely, an identity crisis for everyone who built their career on the other side of that gap.
It's not just strategy
The repricing isn't confined to advertising. Any role where "I know more than you" was the value proposition is being renegotiated, simultaneously. Lawyers doing case research. Journalists building background. Consultants writing category landscapes. Analysts producing summaries. Doctors running differentials. All at the same time.
The signals are already visible. Consulting firms shrinking junior cohorts. Media companies dissolving research desks. Law firms automating associate-level work. Job postings listing "AI-native" as a requirement, not a bonus. CMOs asking why they're paying £50k a month for strategy their internal teams can now produce. The services model -- not individual roles within it, but the entire structure -- is under pressure.
People will tell you AI hallucinates. They'll tell you that you still need a human in the loop, that it can't replace genuine expertise. These things are true. They're also temporary comforts. The trajectory is clear, even if the timeline isn't.
The problems we haven't solved
There's a pipeline problem that nobody is talking about honestly. Junior strategists learned judgement by doing the grunt work first. Hundreds of hours of desk research, competitive audits, consumer interviews. That repetition built intuition over years. If AI does that work now, the rungs at the bottom of the ladder have been removed. How does the next generation develop taste and judgement if the repetition that produced it no longer needs a human?
And what happens to institutions that sell knowing when knowing becomes free? MBAs. Planning certificates. Journalism schools. Law schools. The value proposition of formal education was always partly about access: access to knowledge, access to credentials that signal you'd acquired it. AI doesn't dissolve all of that. But it changes the calculation significantly, and most institutions are not yet reckoning with it seriously.
The instinctive corporate response has been to layer AI onto existing structures. "Use ChatGPT in your workflow." Same roles, same org charts, same briefs, just faster. That's putting a more powerful engine into a horse and cart. The vehicle itself needs rethinking, not just the horsepower.
What remains scarce
Both things are true at the same time: the loss is real, and the progress is also real. The democratisation that nobody mentions enough is genuinely extraordinary. Access to the kind of strategic thinking that used to be the exclusive property of large retainers is now available to anyone with a browser. That isn't a consolation prize. That's a civilisational shift.
What remains scarce, genuinely scarce, is judgement. Taste. Context. The courage to act on uncertain information. The ability to be in a room with someone and make them feel heard. These aren't AI-resistant because they're soft and vague. They're AI-resistant because they compound over time in ways that can't be downloaded, and because they require a track record of decisions, not just access to information.
From person who knows to person who does
If the knowledge part of your job now takes 10% of the time it used to, what fills the rest? Not everyone becomes a builder. Some become editors, curators, orchestrators of AI output. The role doesn't disappear. It shapeshifts. The question isn't "will I have a job?" It's "will I recognise it?"
The transition I've been living over the past year is diagonal, not lateral. I left strategy not because knowing stopped mattering, but because knowing without doing stopped being enough. I've built 20+ live products without writing a single line of code. Not as a party trick. Because the gap between having an idea and testing it collapsed in the same way the gap between having a question and answering it did. And once that happened, there was no reason to stay on the recommendation side of the table.
The new knowledge worker starts with judgement, not research. Uses AI as raw material, not a crutch. Ships, doesn't just recommend. Builds taste through making, not just consuming. And the human skills that got quietly sidelined during the knowledge economy -- persuasion, presence in a room, the ability to sell an idea, the ability to make someone feel heard -- are returning to the centre. They were always part of the job. They're about to become the whole job.
For organisations: stop hiring for knowledge. Start hiring for judgement and output. Rethink the junior pipeline; the apprenticeship problem is real and unsolved. Restructure around what AI can't do, not what it can assist with. Measure people by what they ship, not what they know.
The best strategist I ever worked with wasn't the one who knew the most. It was the one who knew what to do with what they knew.
