Quick-Reference Matrix
Identify the buyer. Read across. Act.
| Buyer archetype | Lead claim | Support claim | Hold in reserve |
|---|---|---|---|
| Founder / CEO | "I ship agent systems under ambiguity" — TinyFish context + Carrier IQ live | "I build 0→1 under hard constraints" — Thermo Fisher, Red Cross | Alibaba scale narrative |
| CPO / VP Product | "My design decisions moved business metrics" — Alibaba mandate + Allē redemption | "I build trust architecture for agentic products" — Trust essay + Carrier IQ | Team-size, org-chart evidence |
| VP / Head of Design | "I hold craft bar and it transfers without me" — Alibaba sprint detail + confidence rule | "I design for agentic systems in production" — Carrier IQ states + Trust essay | TinyFish velocity claims |
| Senior IC panel | "I made this decision, here's why, here's what I'd change" — Carrier IQ interaction states, Alibaba PDP | "I design across consequence levels" — Red Cross intake→payment, Thermo Fisher exception gates | Scale metrics, team narrative |
| Cross-functional Eng / PM | "I diagnose from a trace and change the next release" — Carrier IQ + Thermo Fisher constraints | "I understand what breaks and who absorbs it" — Red Cross 6→1, Trust essay handoffs | Design-leadership framing |
The previous Playbook piece, "Cite the Components, Not the Essays," routed your portfolio by altitude — craft evaluators inspect states, system evaluators inspect controls, leadership evaluators inspect what keeps working after you leave. That routing holds. This piece adds a second layer: which specific claim to lead with, given who's across the table.
The claim is the sentence the evaluator writes in their scorecard after you leave. The same Carrier IQ demo sits in front of a founder and a design panel. The founder writes: "can ship agent systems without organizational cover." The panel writes: "strong interaction-state decisions in the correction flow." The screens are the same. The sentence that gets written depends on which claim you led with.
Founder / CEO
Who this is. Growth-stage or Series A/B. Controls the final decision. May skip structured portfolio review entirely and evaluate through conversation.
What they're scoring. Can this person ship without mature process, staffing, or organizational cover? Designer Fund's founder hiring guide frames the first design hire around autonomy and resourcefulness. The proof standard is demonstrated range under constraint.
Lead with: TinyFish as current-role context — shipped 3 products in 3 months at an enterprise web agent platform, working daily with agent traces and governance. Then show Carrier IQ live. The founder wants to see that you built something real in AI, recently, without a large team behind you.
Support with: Thermo Fisher or Red Cross. Both are 0→1 builds with small teams and measurable adoption. Thermo Fisher's one-designer, two-engineer team building a $20M+-margin platform for six pharma partners is the kind of resource constraint founders recognize immediately.
Cut or hold: Alibaba's $50B GMV narrative. A founder hearing enterprise scale wonders whether you need the infrastructure that scale implies. Keep it for Q&A if they ask about range.
IC+manager tilt: Full IC. Show the work and the decisions, not the org.
Landmine: "3 products in 3 months" invites the quality question. Preempt it: name one specific Carrier IQ decision where you chose the harder path — the confidence rule that renders only from what the system observed, never from the model's self-report. That's a quality decision made under velocity pressure. Show it before they ask.
CPO / VP Product
Who this is. Hiring manager or hiring manager's peer. Appears mid-loop. Controls scope definition and leveling. At companies like Gusto, this person sits in a PM-Engineering-Data-Design leadership quartet.
What they're scoring. Did your design work change the product outcome, and can you explain the mechanism? CZI's interview guide separates impact from product reasoning and individual contribution. Atlassian's design handbook evaluates Product Thinking as a distinct dimension from craft. The proof standard is a causal chain: observation to decision to measured result.
Lead with: Alibaba's mandate story. You identified a structural gap between a consumer-oriented platform and B2B procurement buyers, built the research case, secured executive authority, and led three cross-functional sprints that produced +20% daily transactions and +2.2pt NPS. The CPO hears product reasoning — you named the problem before anyone asked you to, and the numbers moved.
Support with: Allē's dual-surface redesign. The cross-surface product thinking is what this buyer grades for: consumer behavior (3.2× redemption, 47% reactivation) connected to provider economics ($42 CAC) through one design system serving both sides.
Cut or hold: Team-size and org-chart evidence. The CPO is not evaluating your management span. Hold the Trust essay for Q&A — it shows systems thinking but doesn't carry a business metric.
IC+manager tilt: Tilt manager, but through product ownership. The Alibaba mandate story does this naturally — you didn't wait for a brief, you built the case and secured the authority.
Landmine: Metrics without mechanism. The CPO who hears "+20% transactions" immediately asks: what decision caused that? Have the sprint-level answer ready. Which of the three Alibaba sprints drove the largest share of the transaction lift, and what specific design change created the shift.
VP / Head of Design
Who this is. Peer evaluator or hiring manager. Appears mid-to-late loop. At Amplitude, this person expects systems-level thinking and pixel-level craft simultaneously. At Gusto, they want hands-on work on the hardest flows plus the ability to develop senior ICs through direct feedback.
What they're scoring. Can this person hold a quality standard that works when they're not in the room? Amplitude's posting asks for "quality without bottlenecking." The proof standard is a transferable craft artifact: a component contract with named boundaries, a design-system rule another designer applied without you present, a framework you revised against production evidence.
Lead with: Alibaba sprint detail at the component level — the PDP transaction-confidence work, the specific interaction decisions, the design-system components that encoded those decisions for other designers to use. Then connect to the confidence rule from Carrier IQ: an explicit, reusable judgment embedded in a component contract with named failure boundaries and a retirement condition.
Support with: The Trust essay's five handoffs, positioned as a quality framework you apply to agentic work. The two-dials revision — replacing the original trust ladder with autonomy scope and intervention sensitivity — shows you update your own frameworks when production contradicts them. A design leader who revises their model publicly is more credible than one who only publishes finished thinking.
Cut or hold: TinyFish velocity claims. A design executive hearing "3 products in 3 months" at a product role worries about whether the craft held. Keep TinyFish as brief context for why you have production AI depth — nothing more.
IC+manager tilt: Both, explicitly. Show a component or interaction state that demonstrates your craft judgment. Then show how that judgment became a standard someone else could apply. Alibaba's component inventory and Carrier IQ's confidence rule are your two best exhibits.
Landmine: Leadership claims without visible craft. If you lead with "I led the function," this buyer immediately looks for whether you can still do the work. Invert the order. Show the work first, then explain how it became the standard.
Senior IC Design Panel
Who this is. Two to four senior or staff designers. Appear mid-loop, sometimes as a dedicated portfolio deep-dive. Atlassian tells IC candidates to expect detailed questioning about choices. CZI asks for interaction and visual design, iterations, trade-offs, and individual contribution.
What they're scoring. Did you personally make this design decision, was it good, and can you explain the trade-off you rejected? The proof standard is specific interaction states with articulated alternatives. These evaluators are the most likely to open your live portfolio during the conversation and inspect what's actually there.
Lead with: Carrier IQ's interaction states — the structured intake, the parallel-stage visibility, the five-carrier constraint architecture. Walk through one decision where you chose between two viable approaches and explain why. Then move to Alibaba's PDP sprint at the component level.
Support with: Red Cross's intake-to-payment flow across three roles and one record, or Thermo Fisher's exception-handling gates. Both show consequence-aware interaction design where getting the state wrong has real operational cost.
Cut or hold: Scale metrics and team narrative. The design panel is not evaluating your management span or your GMV. Metrics matter only when they explain why a design decision was validated or revised.
IC+manager tilt: Full IC. If you mention team leadership in this room, do it only to clarify contribution — "I designed this; my team designed that" — never to claim credit for work you directed.
Landmine: Velocity claims trigger an immediate quality probe. Lead with one decision in one product, at full resolution. If they ask about speed, explain the constraint that made it necessary and the quality decision you protected despite it.
Cross-Functional Eng / PM Panel
Who this is. PM leads, engineering managers, or technical leads. Appear mid-to-late loop. Atlassian structures this as a triad conversation about trade-offs and collaboration. OpenAI's Engineering Acceleration role requires technical credibility with engineers, data scientists, and researchers.
What they're scoring. Do you understand what constrains us, and does your design work account for those constraints? The proof standard is a causal account connecting a constraint or failure to a design change in the next iteration.
Lead with: Carrier IQ's trace-level work. Walk through a specific diagnosis: an agent trace where the navigation sequence revealed a failure — a wrong field filled, a skipped verification step — and what you changed in the interaction design for the next run. If you can narrate what you saw in a trace, what it meant for the design, and what shipped differently because of it, this room will register technical credibility that most design candidates don't demonstrate. Support with Thermo Fisher's operational constraints: pharma compliance, exception discovery, the approval/rejection controls that exist because the cost of a wrong shipment is a regulatory event.
Support with: The Trust essay's five handoffs, framed as an engineering contract. Intent-setting is input validation. In-progress visibility is observability. The decision gate is a human-in-the-loop checkpoint. Engineers recognize these patterns when you use their vocabulary.
Cut or hold: Design-leadership framing. This room does not care how many designers you managed. They care whether you understand the system well enough to make design decisions that don't create downstream problems.
IC+manager tilt: Full IC with systems vocabulary. Talk in traces and failure modes, not team structure.
Landmine: Framework language without technical specificity. If you say "trust architecture" without explaining what that means at the agent-behavior level, this panel will categorize you as someone who talks about systems but doesn't understand them. Ground every framework claim in one concrete implementation — a specific Carrier IQ stage, a specific Thermo Fisher gate, a specific handoff with the technical mechanism named.
Hybrid Rooms
Most loops mix these buyers. A portfolio deep-dive might include a VP Design and two senior ICs. A final round might put a founder and a CPO in the same conversation.
When the room is mixed, lead with the claim that serves the highest-authority evaluator, then address the secondary buyer's proof standard in your first supporting example. Founder + design-panel room: lead with shipping velocity and autonomy (founder's claim), then immediately show the interaction-state detail at full resolution (panel's proof standard).
When you can't identify the buyer archetype before the conversation starts, the first question they ask tells you. "Who did design report to?" = org-structure doubt. "How involved were you?" = making doubt. "What happened when it went wrong?" = consequence doubt. The question names the buyer. Route accordingly.
- CHI 2026 transparency study: A peer-reviewed study with 12 participants found that eight preferred progressive or on-demand transparency over maximal process visibility, which creates a documented tension with the Trust essay's argument for visible work — worth having a prepared answer when a design panel asks how much of an agent's process to show.
- Anthropic's permission model: Anthropic's agent research frames autonomy as action-specific controls that always-allow, require-approval, or block, reinforcing the two-dials revision over the original trust ladder and giving you engineering-native vocabulary for the cross-functional panel.
- Vercel's design-for-agents practice: Vercel published a detailed account of how they store accepted product decisions as agent-accessible code and require coverage of loading, error, permission, and destructive states — the closest public precedent to the "encoded judgment" claim the VP/Head of Design buyer will probe.
- Amplitude's quality warning: Amplitude's Head of Product Design posting explicitly warns that inexpensive AI-generated UI can erode product quality, making the confidence rule and transferable craft standard directly relevant if that search moves forward.