Interview Protocol

The Advisor Interview

A best-practices guide for capturing how a person thinks, decides, and pushes back — so their judgment can keep advising you.

Format: 1 conversation, ~90–120 min Record: audio + verbatim transcript Draft v1 · for review
Download print version (PDF) ↓

This is a single, semi-structured conversation. Its goal is not a résumé or a list of opinions — it is to capture the reasoning underneath them: how this person weighs a decision, what they reliably doubt, what they actually know cold, and how they say it. The transcript is the raw material; the questions below are a map, not a script.

For the participant

If someone handed you this, here is the honest version of what it is and how it will be used.

I want to keep learning from how you think — even on the days I can't borrow an hour of your time. With your permission, I'd like to record a conversation and use it to build a private, AI-assisted "sounding board" that reflects your perspective, so when I'm working through a hard decision I can pressure-test it against a model of your judgment.

A few promises:
  • It is for my own thinking only. It will never be used to impersonate you, or to speak as you to anyone else.
  • You can skip any question, go off the record at any point, or redact anything afterward.
  • I'll share back what I build if you're curious, and delete the recording whenever you ask.
  • What's most useful to me is your reasoning — not polished conclusions. Half-formed is fine. Disagreeing with the question is better.
Participant name
Consent to record · date

How to run it

Interviewer notes

The quality of the model is set here, not later. A few rules that consistently separate a rich transcript from a thin one:

  • Listen ~80% of the time. Your job is the follow-up, not the speech. Silence earns better answers than the next question does.
  • Don't lead. Never signal the answer you expect — it contaminates the conversation now and the model later. Ask, then wait.
  • Chase stories, not abstractions. "Tell me about a time…" beats "What do you value?" When you hear a principle, ask for the example; when you hear an example, ask for the principle.
  • Follow the energy. Go off-script when they light up. The map is there to return to, not to obey.
  • Capture exact wording. Their phrases and metaphors are their voice — keep them verbatim, don't paraphrase.
  • Mark where the knowledge stops. Note what they defer on. That boundary is what later keeps the model honest instead of confidently wrong.
Why these six parts Parts 1–2 establish story and decision-making; Part 3 captures how they doubt (the most useful and most easily lost signal); Part 4 anchors real expertise so the model can be corrected against facts; Part 5 captures voice; Part 6 sets aside held-out answers you'll use to test the result. Aim for the times shown, but trade freely between them.
Part 1

Story & formation

~15 min
Interviewer Open wide and let them run. Your only moves here are "what happened next?" and "why did that matter to you?"
  1. Walk me through your path. Where did it start, and what are the two or three turning points that made you who you are in your work?
  2. Tell me about an early experience that shaped how you see your field — one that still sits with you.
  3. Who shaped how you think — mentors, rivals, authors? What did each of them teach you?
  4. What's a belief you held strongly early on that you later changed your mind about? What changed it — an event, a person, evidence?
  5. Among people who know your work well, what are you known for?
Part 2

How you decide

~20 min
Interviewer Push for the rule behind the story. After each example: "So what's the general principle there?" Capture the order in which they weigh things.
  1. When you face a hard call in your work, what are you optimizing for — and what are you willing to give up to get it?
  2. Walk me through a recent difficult decision, start to finish. What did you weigh, and in what order?
  3. What's a rule of thumb you trust that many others in your field don't?
  4. When a new idea or proposal lands on your desk, what are the first three things you check?
  5. What's the tradeoff you keep returning to? Speed vs. rigor, cost vs. quality, consensus vs. conviction…
  6. What would you never do, even under real pressure?
Part 3

Convictions & doubts

~20 min
Interviewer · the core This is the most valuable section and the easiest to lose. Capture how they object, not just what they reject — the exact question they'd ask, the tell that tips them off.
  1. What's a widely-held view in your field that you think is wrong, or far too simple?
  2. When someone brings you a plan, what's the first weakness you tend to spot?
  3. What's the question you always ask that other people forget to?
  4. What makes you lose confidence in an idea fast? What's an instant red flag?
  5. Tell me about a time you were the lone skeptic and turned out right. What had you seen that others hadn't?
  6. And a time your doubt was wrong — what did that teach you about when to trust an idea anyway?
  7. If you were reviewing my work and wanted to save me from myself, what would you be watching for?
Part 4

Your domain

~25 min
Interviewer · the anchor Get specifics: numbers, named sources, real examples. When they assert something, ask "where would I go to verify that?" The aim is to separate what they know from what they suspect.
  1. Explain how your field really works to a smart outsider — the model in your head that the textbooks don't quite capture.
  2. What do even experienced people most often get wrong about your domain?
  3. What are the live debates or open questions right now? Where do you stand, and why?
  4. What numbers, benchmarks, or facts do you carry around that anchor your judgment?
  5. Whose work do you trust? What do you read, and what do you deliberately ignore?
  6. What's changed in the last few years that outsiders haven't caught up to yet?
  7. Where does your own expertise run out — what would you hand to someone else, and to whom? This boundary is gold: it's what stops a model from bluffing in your voice.
Part 5

Voice

~10 min
Interviewer You'll capture most of this just by listening across the whole session. These questions surface the rest.
  1. How would your closest colleagues describe the way you communicate?
  2. What are phrases or expressions you catch yourself using a lot?
  3. What kind of language or framing turns you off the moment you hear it?
  4. When you disagree with someone you genuinely respect, how do you say it?
Part 6

Calibration

~10 min
Interviewer · do not skip, do not reuse Record these answers verbatim and set them aside — keep them out of the persona's notes. Later, put the same questions to the finished model cold and compare. This held-out set is your fidelity test: it's the difference between "it sounds like them" and "it predicts them." Prepare 2–3 short, realistic scenarios from your own domain in advance.
  1. Here's a short, realistic scenario from your world — what would you advise, and what's the first thing you'd want to know? [Prepare 2–3 of these.]
  2. I'll describe a plan in two sentences. Tell me the first thing you'd challenge.
  3. Rank these three priorities the way you would, and tell me why the bottom one is last. [Insert three real, competing priorities.]
  4. A quick gut check: [a pick-one question in their domain with a knowable, recordable answer].

Closing

  1. What didn't I ask that I should have?
  2. Who else thinks like you — or usefully unlike you — that I should learn from?

Then close the loop: thank them, restate how the recording will and won't be used, and confirm they're still comfortable. Within a day, send a one-line thanks and the offer to redact anything on reflection.