The head start is getting shorter

A new hire into a commercial team took months, sometimes the best part of a year, to become fully productive. The capability was there on day one.

What wasn't there was everything that lived in other people's heads.

Which accounts to call first. What the pricing exceptions really meant. Why the forecast always slipped in the last week of the quarter.

The people who'd been there a decade carried all of that. It was why they were worth more, and everyone knew it.

Last week I read 2 pieces of research, one after the other, that describe that head start closing. Fast.

Let's dive in.

The first is a field study of just over 5,000 customer support agents. They were given a generative AI assistant and the researchers measured what happened. Productivity - counted as issues resolved per hour - rose 14% on average.

The average hides what matters: novice and low-skilled workers improved by 34%. Experienced, highly skilled workers barely moved (Brynjolfsson, Li and Raymond, NBER working paper, published in the Quarterly Journal of Economics, 2025).

The explanation is the interesting bit. The assistant had absorbed what the best agents did and handed it to the newest ones. It left the experts roughly where they were and pulled everyone else up towards them.

The second is from Stanford's Digital Economy Lab. Employment among US workers aged 22 to 25 in AI-exposed occupations now stands 19% below where it would be had it kept pace with less exposed occupations. The declines are concentrated where AI substitutes for human tasks. Where AI complements people, employment is flat or rising, especially for experienced workers (Brynjolfsson, Chandar and Chen, Stanford Digital Economy Lab, revised August 2026).

You can watch the same mechanism from the other end. A Bloomberg Technology report this week followed laid-off tech workers in Silicon Valley. A few years ago they were hired because they knew a couple of programming languages, which was then about the best skill you could have. Those tasks are largely automated now, and the people who managed them are trying to work out what the market wants from them next.

Different cohort, same mechanism. The scarce thing became the common thing.

Put those together and you get a picture that is very easy to misread.

A Director reads those findings and hears reassurance. The newcomers get the boost. The experienced barely change. Employment holds up for experienced people where AI complements them. Which sounds like: the technology closes the gap from below and leaves you where you are.

Read it from the employer's side.

For years, the premium paid for experience was partly a premium for time. The months it took someone to learn what you know. The stock of context, judgement calls and pattern recognition that only came from the individual’s experience and way of thinking. Organisations paid for it because they had no other way to get it.

Now they have a way to get a good part of it. The support-agent study is that mechanism at small scale: the working knowledge of the best people, extracted, and given to the newest ones on day one.

"Minimal impact on experienced workers" means the tool didn't help you much. It also means the thing you were paid a premium for is being distributed to people who are paid less.

Your productivity held. Your scarcity diminished.

And the mechanism needs no redundancy round to work. It works through the next hiring decision, when the case for one senior hire loses to the case for 2 junior hires plus the system.

It works through the next pay review, when the benchmark for what your experience is worth has moved.

It works through the next restructure, when someone asks what the senior layer does that the tools and the people below it can't...

Seniority puts you at the top of that list.

Back to ramp time.

If I were pricing a senior hire today, I'd split what I used to lump together.

Part of a senior person's value is stored knowledge. How things work here, what's been tried, who to ask. That is now being copied, and the copies are getting cheaper every month.

The other part is judgement under conditions nobody has documented. The customer call that is going wrong in a way the playbook doesn't plan (yet). The forecast you don't believe. The restructure where the org chart says one thing and the people say another. That part hasn't been copied. The Stanford data says the market is still holding on to the people who have it.

The problem for most senior professionals is that they have never separated the two. In your own head they are both "experience" hard to replace.

So separate them. This week.

Take your last 12 months and list the moments where you made a decision that another capable person, with the same information and the same tools, would have got wrong. Decisions, rather than tasks. Most people reach 5 or 6 and then slow down. Those 5 or 6 are what is still scarce.

Then the value question:

Who else would pay for those decisions? A smaller company that can't afford them full-time. A board. A fractional seat across 2 or 3 businesses. A consultancy.

If you can't answer that, you've priced yourself the way I used to price hires, and the head start you're relying on is shrinking while you’re reading this.

That's the work Zelova is built for. The guide takes what you actually know how to do, separates what's being copied from what isn't, and works out who else would pay for the latter. Then it turns that into a plan with more than one outcome in it, helping you build optionality in yoru career.

It's a career strategy service, still being built. If any of this resonates, the waitlist is open at zelova.ai.

When someone new joins your team now, how long before they're doing most of what you do? Be honest about it. And has that number moved in the last 2 years?

Reply and tell me. I'd like to hear your version.

David