A predictable interview question is doing two jobs at once. It raises a concern about your background that needs an answer — the individual dossiers in this section handle that. It also discloses something about the company. The question got asked because of whatever is broken, missing, or contested inside that organization right now, and the shape of the question tells you which.
This is a method for reading the second job while you answer the first.
There is also a behavioral effect, and I'd label it moderate confidence because I've watched it more than I've measured it. A candidate braced to defend paces differently than a candidate listening for what the question exposes. It shows up in follow-ups, in how quickly you fill a pause, in whether your answer sounds rehearsed. Interviewers register that difference without being able to name it, and they register it whether or not the concern itself clears.
What each question likely reveals
"Are you hands-on?" — Two opposite conditions produce this question. The company needs senior craft judgment close to ambiguous product problems and wants to know you won't retreat into process. Or the company is understaffed and needs the design leader to be the primary production designer. The words are identical; the jobs are not. (High confidence. I'll walk this one through the full framework below.)
"Can you move fast?" — A design leader or design process slowed product development here, and it left a mark. The concern is rarely about your personal speed. It's about whether you'll add process overhead that engineering and product will feel as friction. Underneath that: how much organizational authority design currently holds. A team that has to ask permission to slow down has never been trusted to set its own pace. (High confidence.)
"Why product? You've been in consulting." — Someone on the committee carries a prior about consultants — that they optimize for the presentation over the shipped outcome, or that they leave once the work gets operationally boring. Either this company hired a consultant-to-product convert and it went badly, or one influential panel member has a pattern-match they'll apply regardless of what you say. The underlying worry is staying power and ownership. (Moderate confidence — the specific scar varies by company, and you usually can't see it from outside.)
"What's your AI experience?" — This one splits by company type. At an AI-native product company, it's a legitimate scope check: can you design for probabilistic systems and interaction patterns that don't have established conventions yet? At a traditional product org that recently shipped AI features, the question often reveals that nobody internally agrees on what AI design capability means. Your answer will be graded by people who may not be able to tell real capability from buzzword fluency. Figure out which room you're in before you answer. (Moderate confidence.)
"Have you managed teams this size?" — This is a recovery hire. The usual pattern: design headcount grew faster than leadership capacity, the previous leader couldn't operate across the added management layers, and the team has been running thin on senior coverage since. The committee is anxious about repeating one specific failure — management span, cross-functional influence, hiring judgment, org design — and you need to surface which one before you can answer anything but the generic version. The scaling dossier covers the response itself. (High confidence.)
Five activation rules
Timing matters as much as content. Match the mode to the situation.
Mode 1: Preempt in your opening materials. For concerns so predictable that letting them surface in conversation means the interviewer has already spent ten minutes building a negative hypothesis. The consulting-to-product bridge belongs here. Your positioning narrative should resolve it before anyone thinks to ask. Issue #8 covered why the recruiter screen is where the two-sentence version has to land.
Mode 2: Raise it yourself during qualification. For questions that are really about the company's operating condition, where you need the answer to decide whether the role is right. The hands-on question usually belongs here. Ask about team size, staffing plans, and the expected split between leadership time and production time before they ask whether you can do both.
Mode 3: Answer directly only when asked. The default for concerns that are real but not disqualifying. AI experience usually sits here. Don't volunteer a defense against a concern nobody has activated — the materiality rule from Issue #9 still holds.
Mode 4: Repair when the concern has already shaped their read. Use this when the interviewer's framing tells you they've formed a view. Specificity is the tell: "I noticed your last three roles were consulting engagements" is a hypothesis being tested, not an open question. Repair means naming the pattern they've spotted before you redirect to the evidence that complicates it. Concede the fact, contest the inference.
Mode 5: Treat the underlying condition as an exit signal. For when what the question reveals is disqualifying. If "Are you hands-on?" turns out to mean "you will be our only designer," and a one-person design function isn't the job you want, no answer improves your situation. What you need is the information, and you now have it.
The hands-on question, start to finish
Three words that can describe roles with almost nothing in common. Two current postings show the range.
Amplitude's Head of Product Design posting uses "player-coach" for a leader accountable for 15 designers. The hands-on component there is craft judgment — pairing with designers on hard UX problems, giving feedback, holding coherence across product surfaces. The leader doesn't own any single surface end to end.
Fin's Head of Design posting uses "hands-on" for the company's only full-time designer. The scope covers product UI, design systems, and brand. The title says Head of Design, but the scope is a senior IC role with no team.
Activation rule: Mode 2. Raise it during qualification. Ask about team size, current staffing, and how the role divides leadership time from production time — Issue #9's qualification prompts have the specific wording. You need the answer to evaluate the role, and asking early signals that you know these are different jobs, which is itself a piece of evidence about the craft judgment they're trying to assess.
If they ask first, answer with work. Thermo Fisher and Red Cross demonstrate hands-on craft judgment applied to ambiguous product problems. Alibaba demonstrates organizational scale. Which one leads depends on which version of "hands-on" they mean, and you know that by now — either because you asked, or because you read the posting through this lens before you walked in.
After you respond. If the interviewer moves to your process on a specific project, or asks how you'd approach a product problem they're currently stuck on, the concern has probably cleared. If they rephrase — "But would you actually be in Figma?" or "How much of your week would be hands-on versus managing?" — it hasn't. Change the evidence rather than repeating the same framing with better polish. Offer a specific artifact, or a concrete decision you made inside the work. You can also ask what their previous design leader's hands-on involvement looked like; that question tends to surface the real concern faster than another round of your evidence. Ask it after you've answered, though. Not instead.
The exit check. If "hands-on" means solo designer and that isn't the job you want, say so. Withdrawing cleanly from a mismatched role costs you nothing.
Next-question tracking
What the interviewer asks next is your best signal for whether the answer worked.
When the same concern comes back — rephrased, restated, approached from a new angle — your answer didn't settle it. Moderate confidence on this reading. Work out what they're still testing, and change what you're offering as proof.
When they advance to craft, scope, or operating questions, call it provisional clearance rather than confirmation. In a structured interview format, the topic shift may just be the next item on a predetermined list. The signal gets stronger when the interviewer builds explicitly from something you said: "You mentioned the Thermo Fisher redesign — how did you handle X there?" carries more information than a generic pivot.
Log it. Which evidence you used, what they asked next, whether the concern came back later in the process or from a different interviewer. Patterns across four conversations are worth more than your read on any single exchange.
Answer first, then ask
When a hard question exposes something about the company's condition, you'll want to ask about it immediately. Don't. Answer fully first, then ask your diagnostic question from real curiosity about the role.
The worked example above marks the moment precisely: you give your evidence on hands-on craft judgment, and only then ask what the previous leader's involvement looked like. That question, arriving after a substantive answer, is a senior leader doing due diligence. The same question arriving before you've given them anything sounds like someone stalling.
Your evidence first, your diagnostic question second. This holds across every objection in this section.
Quick reference
Scan this before a conversation.
"Are you hands-on?"
- Reveals: Team staffing level; whether the role is craft-judgment leadership or IC production with a leadership title. (High confidence.)
- Default mode: Raise during qualification (mode 2).
- Cleared when: Interviewer moves to specific craft or product-problem questions. Still active if: they rephrase around time allocation or tool usage.
"Can you move fast?"
- Reveals: A previous design leader or process created friction with product and engineering. Design may lack organizational authority. (High confidence.)
- Default mode: Answer directly when asked (mode 3). Lead with compressed-timeline evidence — Red Cross national deployment in six months, Equinox+ MVP in three months. TinyFish's three-month platform launch works as verbal support for recent speed under ambiguity. Escalate to mode 2 if early signals suggest design has no structural authority to set its own pace; at that point it's a qualification question about the role, not an objection to answer.
- Cleared when: Interviewer asks about your process on a specific project. Still active if: they keep returning to timelines or velocity in the abstract.
"Why product? You've been in consulting."
- Reveals: A bad prior experience with a consultant-to-product hire, or an influential panel member with a fixed pattern-match. (Moderate confidence.)
- Default mode: Preempt in opening materials (mode 1).
- Cleared when: Interviewer stops referencing consulting and asks about shipped product outcomes. Still active if: they keep circling tenure or commitment.
"What's your AI experience?"
- Reveals: Either a legitimate scope check (AI-native company) or internal confusion about what AI design capability means (traditional company adding AI features). (Moderate confidence.)
- Default mode: Answer directly when asked (mode 3).
- Cleared when: Interviewer asks about specific AI design challenges — model behavior, probabilistic UI, evaluation. Still active if: they ask vague follow-ups about "AI thinking." The company may not know what it needs.
"Have you managed teams this size?"
- Reveals: Recovery hire. Design headcount grew faster than leadership capacity, and the committee is anxious about repeating one specific failure — management span, cross-functional influence, org design. (High confidence.)
- Default mode: Answer directly when asked (mode 3), using the scaling dossier's evidence sequence.
- Cleared when: Interviewer shifts to how you'd structure the team or handle a specific org-design problem. Still active if: they restate with different numbers — change the evidence you offer.
- Amplitude's player-coach scope: Their Head of Product Design posting is one of the clearest current examples of "hands-on" meaning craft judgment across 15 designers rather than personal production — worth tracking as a reference point for how staffed orgs describe the role.
- Structured probing changes outcomes: A 2025 experiment found that adding structured probes to asynchronous video interviews improved both perceived opportunity to perform and interviewer ratings, which means a topic shift in a structured format may reflect the script rather than a cleared concern.
- Anthropic's authorship boundary: Their candidate AI-guidance policy encourages AI use for application prep but prohibits it in assessments unless explicitly allowed, drawing a line between AI fluency and independent judgment that maps directly to how "What's your AI experience?" splits by company type.
- Candor timing is conditional: A field study found that candidly disclosing negative information could improve offer likelihood, but a separate experiment showed perceived competition can make disclosure backfire — supporting the activation rules over a blanket preemption strategy.

