← All resources
Resources

How to Choose an AI Coach (the Human Kind)

Let's clear up the confusing bit first: an "AI coach", as we use the term, is a real human who teaches you AI skills — not an AI that coaches you. So before you choose an AI coach, know which kind you're shopping for. Plenty of apps sell chatbot mentoring; this guide is about the person who gets on a call with you, watches how you actually work, and makes you measurably better with these tools.

The trouble is that "AI expert" is currently the cheapest title on the internet. It costs nothing to claim, and demand is high enough that plenty of people are coaching on six months of enthusiasm. If you're about to pay someone $40–75 an hour to teach you something that could reshape your working week, you deserve a better filter than a confident LinkedIn bio.

This is that filter — the checklist we'd hand a friend: what verified skill looks like, how to match a specialty to your problem, which red flags should end a conversation, and how to get real signal from a free intro. If you'd rather browse while you read, the coach directory shows everything discussed below: scores, specialties, reviews and rates.

Verified skill: why an AISA score beats a bio

Start with the one question that disqualifies quickly: can this person prove they're good at AI, or do they merely say so? Self-declared expertise is unfalsifiable — which is precisely why it's everywhere.

Every coach on humanscoach has scored 80 or above on the AISA assessment, a standardised test of practical AI skills; a score of 90+ earns Expert tier. The score doesn't tell you everything — it can't measure patience or teaching instinct — but it draws a hard floor under competence. Treat it like a personal trainer's qualification: it doesn't guarantee they're the right trainer for you, but it does guarantee the technique they're teaching isn't improvised. A bio guarantees nothing at all.

Match the specialty to your actual problem — then check the receipts

A brilliant prompting coach can still be the wrong hire. Coaches specialise — in prompting, automation, strategy, writing, or working with complete beginners — and the fit between specialty and problem matters more than overall brilliance.

Picture a Head of Ops whose Monday morning disappears into copy-pasted reports. She needs an automation coach who has wired these tools together before, not someone whose real strength is prompt technique. Or a novelist who wants AI for research and structure but worries it will sand the character out of her sentences — she needs a writing coach whose whole practice is protecting a voice, not replacing one. Name your outcome first; let the specialty follow.

Once the specialty fits, check the receipts. Reviews from past clients tell you whether learning actually happened. A portfolio — automations built, workflows shipped — shows the coach has done the thing rather than talked about it. "What have you made with this?" is a fair question to ask anyone charging for expertise.

Coaching style is the part everyone forgets

Two coaches with identical scores can be completely different to learn from. Some set homework and check it like a stern piano teacher; others explore alongside you and let the session follow your curiosity. Neither is better — but one of them is better for you.

Sessions happen over Zoom or Google Meet, and a good coach sets practice between them — so be honest with yourself about which kind of accountability you respond to. If a stack of exercises makes you feel behind rather than motivated, a gentler, project-led coach will take you further than a rigorous one whose sessions you start quietly avoiding.

Red flags worth ending the call over

Some patterns reliably predict a poor investment:

  • The hype merchant. Everything is revolutionary, nothing has limits, and they cannot name a single thing AI does badly in your field. Coaching requires honesty about the tool's edges; someone selling a revolution can't afford any.
  • The tool-of-the-week collector. Their feed is a parade of app launches. You'd come away knowing product names, not the durable skills — framing a problem, prompting well, verifying output — that survive the next model release.
  • Nothing verifiable. No assessment score, no portfolio, no reviews — just certainty. Confidence is not a credential.
  • They talk in keynotes. Ask a practical question and get an answer about AI's impact on your industry. You need someone who can tell you what to do on Tuesday.

Five questions to ask in your free intro session

Every coach offers a free 30-minute intro session, and most people spend it on pleasant small talk. Treat it as an interview instead:

  1. "What have you built with AI in the last month?" You want recent and concrete. A good answer sounds like a specific project; a bad one sounds like a talk they give.
  2. "Here's my actual problem — how would you start?" Bring a real one: the Monday report, the proposal backlog. Watch whether you get a specific approach or generic enthusiasm.
  3. "What would you have me practise between sessions?" Good coaches design practice. If the answer is essentially "book more sessions", that tells you the model.
  4. "What is AI still bad at in my field?" The honesty test. A trustworthy coach answers immediately and specifically. A hype merchant flounders.
  5. "How will we know it's working after three sessions?" You're listening for an outcome — a report that builds itself, a first draft in half the time — not a feeling.

If a coach handles all five well, book them. You're not gambling much either way — the first paid session carries a money-back guarantee.

A note on money

Pricing is per session, with no subscription. Coaches set their own rates — typically $40–75 an hour, with executive advisers and deep specialists at $100 and up. The full breakdown is on the pricing section.

Per-session pricing keeps the incentive honest: a coach only keeps earning by staying useful. Resist the temptation to sort by cheapest — a low rate teaching the wrong things is the most expensive option on the list. And if you've read this far thinking "I could pass that assessment myself" — you may be on the wrong side of the marketplace. We're always looking for coaches.

Frequently asked

Isn't an "AI coach" just an AI?

Often, elsewhere — the phrase usually describes software that coaches people. Here it means the reverse: a vetted human professional who teaches you AI skills one-to-one over Zoom or Google Meet. Every coach on humanscoach is a real person who has scored 80+ on the AISA skills assessment.

What is an AISA score, and can I check it myself?

AISA (aisa.to) is a standardised assessment of practical AI skills — a scored test rather than a self-description. Coaches must score 80 or above to join the platform; 90+ earns Expert tier. Each coach's score is shown on their profile in the directory, so you never have to take competence on faith.

What if I choose the wrong coach?

The system is built for low-stakes trying. The first 30-minute intro with any coach is free, and your first paid session carries a money-back guarantee — a full refund, or a free session with a different coach. Because pricing is per session rather than subscription, switching costs you nothing.

Couldn't I just learn from ChatGPT itself, or from YouTube?

You can learn a fair amount that way — AI is a remarkably patient tutor. What it can't do is watch how you actually work, catch the habits you can't see, or care whether you did the practice. Courses have the same gap: they teach about AI in general, and finishing them is on you. A coach compresses months of solo trial and error into a few observed, personalised hours.

Meet a human coach — free.

Every coach offers a free 30-minute intro. No card, no commitment.

Browse coaches →