Ten people whose living depends on knowing what machines cannot do, asked to say what that is.
Everything upstream of the shortlist is already gone. Mapping a market used to take a researcher three weeks and now it takes an afternoon. It hasn’t been access for years: everybody can find everybody. What clients pay for is judgement and the willingness to be wrong in public.
The mapping is done. In regulated markets every institution can produce the same list of forty names, so the list is worth nothing. What a machine cannot tell you is who the regulator trusts. That sits in a hundred conversations that were never written down.
The bit the machines do was never the hard bit in my world. The hard bit is knowing which plant director actually did the turnaround and which one stood near it. The research got cheaper and the judgement got dearer. Suits me fine.
In life sciences the bottleneck was never finding people. Everyone knows everyone; the science community is a village with a conference calendar. AI has made the long list free and the short list more valuable.
My market builds the tools, so we felt it first. Sourcing is automated, outreach is automated, and candidates know it. The paradox is that automation made the human call the premium product.
At board level the work was never finding names; there are only so many people who have chaired an audit committee through a restatement. What is left is the part that was always the job: judgement, discretion, and telling a chairman something he does not want to hear.
Consumer got hit twice: the tools changed our work and changed our candidates’ jobs at the same time. The leaders who survive it are the ones who stopped pretending to be the smartest analyst in the building.
The professions bill time, and the machines are eating the billable hour from the bottom. The people I place are having to answer a question they have deferred for twenty years: what exactly are we charging for?
Increasingly the AI is the transformation, and my people are the ones sent in to make it real inside businesses that bought the licence and got nothing. The gap between the deck and the shop floor has never been wider.
Energy runs on engineers, and engineering is where the tools genuinely deliver. A project director used to be the person with the answers; now they are the person who decides which answers to trust.
Six of the ten said the research layer of their own profession is already automated, and not one of them mourned it. The consistent claim was that the lists were never the product: the market map every company can now generate in an afternoon was, in their telling, always the cheapest part of the work.
The split is in where the change lands. Six put it in their own work: judgement, discretion and the willingness to hold an unpopular view are now the whole of what a client is buying. Four put the bigger change on the people they place: leaders who ran functions by instinct being asked to run them by model, and public words that are now searchable forever.
What nobody said may be the most striking finding. Not one of the ten described the technology as a threat to their own business, and not one predicted a smaller industry. Whether that is clear sight or the view from inside is a question the next volume can test, because every answer here is dated and kept.
The same question, answered in writing by each of the Twenty.
The tools flattened the field between the global companies and the rest of us. Nobody misses the database advantage.
Candidates use it too. The polished CV is free now, which makes the conversation matter more.
Speed. A three-week long-list takes a day. Clients now expect the judgement at the same speed, which is the real pressure.
The people we place are being asked to lead functions they can no longer personally do. Most were not trained for that.
Fees will follow the value, and the value now sits at the start and the end of the process. The middle has quietly disappeared.
Less than everyone claims. Senior appointments were never won on research. They are won on trust, and there is no model for that.
The typing got faster. The listening didn’t.
The routine parts of my work went first, which returned me to the part I started for: the conversations.
Our predictions got better and our excuses got worse. Both are progress.
Committee papers are drafted by machine now. Judgement of people, thankfully, is not.
The translation layer went. What remains is knowing which markets a reputation actually travels to.
Founders now arrive with a machine-made shortlist and pay me to disagree with it. Good trade.
The paperwork evaporated. The pit visits remain. Right things went.
The brief-writing got instant. Knowing which creative leader can actually rebuild a brand did not.
I still do the same job: telephone, kettle, honesty. The research my grandson automates.
Preparation became free, so preparation stopped impressing. Presence cannot be generated.
Reconciliation died quietly. Interpretation got a pay rise. My candidates are fine.
My market invented the disruption and is now surprised to be disrupted. The humility is overdue and useful.
Public appointments now arrive pre-researched by everyone. Transparency raised the floor and my blood pressure.
The gatekeeping got weaker, which is the best thing the machines have done. Talent shows up in data it never showed in rooms.