Mary Fung
essayJune 20, 2026

Hiring for the AI age

The hiring signal is not who knows the newest tool. It is who can learn, judge, translate, and own the result.

"Knows AI tools" is a weak hiring signal.

Tools change too quickly. Tool fluency is easy to exaggerate. A person can sound current and still have no judgment. Another person can be behind on the latest interface and still learn faster than the rest of the team once the work requires it.

The better hiring signals are behavioral.

Look for agency. Ask when the person made progress without a clear path. Listen for whether they waited, escalated, complained, or built a first map.

Look for learning velocity. Ask what part of their workflow changed in the last six months. A vague answer is useful data.

Look for judgment. Give them two plausible AI outputs and ask which one they would ship. The answer matters less than the reasoning.

Look for taste. Ask about bad work they have shipped. The best people can name what was off without becoming defensive.

Look for accountability. Ask what they do when AI gives them an answer they want to be true. The right answer includes checking, not vibes.

Look for translation. Ask them to explain a technical constraint to a business leader, or a business requirement to a technical team. AI-enabled teams need people who can move between those languages without flattening the meaning.

Curiosity matters too, but curiosity is easy to perform. The useful version of curiosity leaves evidence: experiments, changed workflows, better questions, shared notes, faster learning, sharper output.

Different roles will show these traits differently. Builders should be able to make things real and explain the tradeoffs. Operators should know where the workflow breaks. Marketers should understand audience, message, and proof. Designers should know how users actually move. Managers should know how to set standards and inspect quality.

The interview should not test whether someone can name tools.

It should test whether they can use a tool to improve a piece of real work without losing responsibility for the result.

Good question: show me a workflow you changed because AI made the old version unnecessary.

Better question: what did you stop doing after the new workflow worked?

← back to the field