For: Founders building a team that will be smaller and denser than the last one

Who do I actually need to hire in the AI era?

Where the line actually sits

Half

The headcount to design the role around

Write the job for a team half the size of the org chart you inherited from companies built before any of this. If the role only makes sense at the old size, you're hiring for a company that no longer exists.

3 outcomes

In the first 90 days, written before the job post

Measurable and owned. If you can't write them, you aren't ready to hire — you're hoping someone arrives and tells you what they should be.

90 minutes

Live exercise instead of a résumé conversation

Give them a real problem from your business. Watch what they hand to a machine, what they keep, and whether they check the output. That hour tells you more than three interviews.

Day 90

When you decide, out loud

The expensive mistake isn't hiring wrong. It's knowing by week six and acting at month ten while the rest of the team watches you not act.

What changed in the last year

• The junior rung thinned out. Work that used to be given to a first-year analyst to build judgment is now done in minutes, which is efficient this year and a talent-pipeline problem within five. Hire some juniors anyway, deliberately, and know why you're doing it.
• AI-fluent and AI-native became different candidates. Fluent people use the tools. Native people rebuild the work so fewer steps exist. Ask what they stopped doing entirely last year — fluent candidates list tools, native candidates describe a process that no longer exists.
• Contract-to-hire lost its stigma. Paying for the first 90 days as a project is now normal enough that good people accept it, and it reveals more in three weeks than any reference call.

What most founders get wrong

• “We'll hire a team and figure out structure later.”

  Structure arrives anyway, just accidentally, and undoing an accidental org is far more painful than designing one.

• “This candidate is great with AI.”

  That's a floor, not a differentiator, the way being great at email was in 2005. What matters is what they removed from the process, not which tools they opened.

• “We can't afford someone senior.”

  On a small team you can't afford someone junior in a role that requires calls. One senior person and better tooling beats three people who all need direction from you.

When I'd tell you the opposite

• The work is genuinely volume work with a clear right answer and a human has to be accountable for it — then you do need hands, and you should say so plainly rather than dress it as a leadership role.
• You're deliberately building a training pipeline. Hiring juniors is then a strategic cost, not an efficiency mistake, and worth defending.
• A regulated function where the standard requires named, qualified humans. The org chart is set by the regulator, not by you.

Related questions

How do I tell if someone is AI-native in an interview?

Ask what they stopped doing entirely in the last year. Fluent candidates name tools they added; native candidates describe work that no longer exists.

Should I still hire junior people?

Yes, but on purpose and with a plan for what they learn. The economics say skip them; in five years the companies that skipped them will have no bench.

What's the right first executive hire?

Whichever function will break first at three times your current volume. Not the function you enjoy least — that's a delegation decision, not a hiring one.

Is contract-to-hire fair to candidates?

It is when it's paid properly, scoped clearly, and both sides can walk. It protects them from a bad fit as much as it protects you.