Four Assets, Four Tiers, Sixteen Ways to Get the Story Right or Wrong

The next time a role goes live, you won't have time to figure out which version of your story to tell which room. H2 headcount unlocks mid-July. Postings will surface while people are still clearing their holiday inboxes. This piece is the reference you read before that happens.

You have four published assets. Four target tiers. Sixteen intersections where an asset either proves the right thing to the right buyer or proves the wrong thing and costs you the room. The evaluator will not do the translation work for you. A pharma supply chain case that reads as consequence literacy to a healthcare panel reads as irrelevant domain experience to an AI-native one, unless you know which sub-asset to pull forward and which narrative to leave behind.

What follows is organized by asset, not by tier, because you need to understand what each piece of work can prove before you decide what it should prove in a given conversation. After the four asset sections, a set of tier cards gives you the deployment order for each audience. Print the card. Read it before the call.

"Trust Is the New Interface"

The essay is roughly 1,900 words of published thinking that names a specific design problem: agentic systems move the design surface from deterministic path optimization to intent communication, reliability signaling, handoff architecture, and incremental trust. It proposes a five-handoff framework (Intent-Setting, In-Progress, Output Review, Decision Gate, Loop Feedback) and a three-state trust ladder (watch, verify, delegate). It references Alibaba and Thermo Fisher as contexts but presents no shipped AI metrics.

The essay proves judgment without proving execution. That gap determines how it functions at every tier.

AI-native (primary proof of category judgment, high confidence). AI-native evaluators listen for architectural and systems-level language. They want to hear you name the structural problems their teams work on daily, using vocabulary that proves you've internalized the material. Lead with the essay as a threshold credential. Say "watch, verify, delegate" in the first two minutes. Say "I've mapped the five handoff points where agentic workflows break trust, and I've built against them." Then pivot to the labs immediately. The essay gets you into the conversation. The labs carry it.

Misframing risk: dwelling here. If you spend more than ninety seconds on the essay without moving to Agentic Labs, the evaluator codes you as a theorist, and theorists don't get hired at this tier.

Growth-stage regulated (supporting evidence, high confidence). Growth-stage regulated evaluators listen for velocity-under-constraint language. They need proof you can move fast without breaking things that carry real weight. The Decision Gate section explicitly names QA regulatory release as an irreplaceable human gate, using Thermo Fisher as the example. That single passage does more work for regulated audiences than the rest of the essay combined. Pull it forward. Skip the full framework. Go straight to the Decision Gate, name the regulatory constraint, then pivot to Thermo Fisher. The claim: "I've already published my thinking on where agent autonomy ends and human sign-off is non-negotiable."

Misframing risk: leading with the essay before the Thermo Fisher case. Regulated audiences need consequence proof before they'll credit your AI thinking. The essay earns attention only after you've established that you've shipped under real regulatory weight.

Enterprise platform (supplementary context, high confidence). Enterprise evaluators listen for organizational-influence language: consensus, cross-functional alignment, moving complex systems without owning them. The essay addresses none of that. Use it only if the conversation turns to AI strategy and you need to demonstrate a published point of view. One sentence: "I published a framework for designing trust in agentic systems. Happy to share the link." Do not walk through it.

Healthcare/vertical SaaS (supporting evidence, but only the Decision Gate passage, high confidence). Healthcare evaluators listen for domain-compliance language. They need to hear that you treat regulatory constraints as design material. The Decision Gate passage delivers: "QA regulatory release as a required human gate" is domain-native phrasing a healthcare panel will recognize immediately. Use it as a bridge from Thermo Fisher to your AI thinking. Never as a standalone asset. And never open with the consumer-facing insurance-quote anecdote from the essay's introduction. Healthcare evaluators will hear "consumer AI" and mentally file you in the wrong category before you've started.

Alibaba

The case presents a B2B platform redesign at $50B+ GMV scale: +7% DAU, +20% transactions, -47% security concerns. You were Head of Design & Research, North America. The narrative runs from diagnostic research (32 cross-functional interviews, Baymard audit, gaze tracking on $500+ order sessions) through three sprints addressing first impression, search, and transaction confidence. The full case includes an agentic rebuild vision describing a natural-language procurement brief that becomes structured parameters, agent-driven supplier qualification, and autonomous transaction completion.

Two assets live inside this case. The shipped platform story and the agentic rebuild vision. They serve different tiers and should almost never be presented together at full length.

Enterprise platform (primary proof of scale and organizational altitude, high confidence). Enterprise evaluators are testing whether you can move a complex organization without owning it. Foreground the diagnostic method. Say "32 cross-functional interviews before we touched a pixel." Say "desktop was 25% of traffic but 80% of transaction value, so we had to convince the org to prioritize the surface that looked smaller but carried the revenue." That sentence proves organizational influence, which is the currency in enterprise rooms. The agentic rebuild vision is a closing move only. Use it if the interviewer asks about your AI perspective, framed as: "Here's how I'd rebuild a system I already understand deeply." Save it for the end of the conversation.

Misframing risk: opening with the agentic vision. Enterprise evaluators will hear "speculative redesign" and discount the shipped work.

AI-native (supporting evidence via the agentic vision, high confidence). The guideline's starting position was supplementary, but the agentic vision earns a stronger gate here. It shows a procurement workflow where an agent parses a natural-language brief, traverses search, evaluates trust signals against the brief, and completes transactions autonomously. That is a complete agentic loop grounded in a real domain you've already shipped in. Foreground the vision. Subordinate the shipped metrics. The claim: "I redesigned B2B procurement at $50B scale, and I've already mapped how an agentic architecture would replace the deterministic flows I built."

Misframing risk: walking through the sprint-by-sprint shipped narrative. AI-native rooms will lose patience with conversion-rate improvements on traditional web surfaces. They want the agentic thinking. Get there in under a minute.

Growth-stage regulated (supplementary context, high confidence). One sentence to establish altitude: "I led design for Alibaba's B2B platform at $50B+ GMV." Then move to Thermo Fisher. Growth-stage regulated rooms are evaluating consequence literacy and speed. Alibaba proves scale but lacks the constraint-dissolution evidence they're hiring for.

Healthcare/vertical SaaS (supplementary context, high confidence). One sentence for altitude, then pivot. Healthcare evaluators will not see B2B e-commerce as domain-relevant. The bridge to their world is too long. Say "I've led platform design at $50B+ GMV scale" and move directly to Thermo Fisher, which speaks their language natively.

Thermo Fisher

The case presents a 0-to-1 pharma supply chain platform covering six pharma partners, nine manufacturing sites, three continents. Published outcomes: $20M margin opportunity, 100% partner adoption, 83% IRR, 42% overhead reduction. The design surface is exception-first: orders, batches, dashboards, forecasts, and capacity modules all organized around surfacing problems before they reach delivery gates. The case names Good Distribution Practice and current Good Manufacturing Practice as regulatory constraints shaping the design. The full case includes a five-agent agentic vision where each module becomes a continuously running agent, with one explicit human gate: batch QA release, where regulatory signature is irreplaceable.

This asset does the most work across the most tiers. It is also the one where misframing costs the most, because the shipped platform and the five-agent vision serve fundamentally different evaluation frames.

Healthcare/vertical SaaS (primary proof of consequence literacy, high confidence). Lead with this. Healthcare evaluators need to hear domain-compliance language before they'll listen to anything else. The claim is specific: "I designed a pharma supply chain platform where a visibility gap was a compliance failure, and I embedded GDP/cGMP constraints into the workflow so they accelerated operations instead of gating them." Exception-first design became the default view, so the regulatory requirement became the interaction pattern, woven into the product rather than layered on top. Healthcare evaluators will recognize this as the hardest design problem in their domain.

Introduce the five-agent vision only after you've earned trust with the shipped platform. Frame it as: "Here's how I'd extend this with five continuously running agents, keeping QA release as a required human gate." That last phrase is load-bearing. Never omit it with this audience.

Misframing risk: opening with agents. Healthcare audiences hear "automation in pharma" and their compliance instinct activates against you before you've demonstrated that you understand why it should.

Growth-stage regulated (primary proof of consequence literacy and constraint dissolution, high confidence). Same core claim as healthcare, with one critical addition: the timeline. "Twelve months, 0-to-1, six partners adopted." Growth-stage companies are evaluating speed alongside consequence literacy. That sentence proves you can ship under regulatory weight at a pace that doesn't terrify a board. The 0-to-1 framing matters more than the nine-sites framing here. You built this from nothing. That is the growth-stage signal.

Misframing risk: overemphasizing enterprise scale (nine sites, three continents) in a way that makes a growth-stage company worry you'll import heavyweight process. Lead with the build, not the footprint.

AI-native (supporting evidence via the five-agent vision, high confidence). The five-agent architecture is the strongest autonomy-threshold evidence in your portfolio. Five agents running continuously across orders, batches, dashboards, forecasts, and capacity. One irreplaceable human gate at batch QA release. This is the specific design judgment AI-native companies need and almost no candidates can demonstrate from real domain work. Name the agents. Name the human gate. The claim: "I've already designed the autonomy boundary for a five-agent system in pharma, where getting the threshold wrong has regulatory consequences."

Misframing risk: telling the full Thermo Fisher story before reaching the agentic vision. AI-native rooms are testing architectural fluency, not supply chain expertise. Open with: "I built a pharma platform that I've since redesigned as a five-agent architecture," and go straight to the agent logic. The supply chain narrative is background. The agent logic is what they're evaluating.

Enterprise platform (supporting evidence, high confidence). Enterprise evaluators will credit the partner adoption (100%, six pharma companies) and cross-organizational complexity. Use it to prove you can build consensus across organizational boundaries with external stakeholders who have no obligation to adopt your platform. Subordinate the regulatory detail unless the enterprise role involves regulated domains.

Agentic Labs

The labs are live, functional agents you built solo. Carrier IQ runs parallel insurance carrier quote flows with broker-side evaluation. UAT Sentinel turns plain-English test goals into browser execution with failure flags and release recommendations. Brand Pulse produces live brand-health scores from Reddit and X. Retail Velocity scans restaurant ordering pages and creates ranked account lists. Creator Scout scans YouTube and TikTok for ranked creator shortlists.

No adoption metrics. No revenue. No team. These are applied research, and the positioning must say so explicitly every time, or the evaluator will measure them against production standards they were never built to meet.

AI-native (primary proof of hands-on building, high confidence). The guideline's starting position was "supporting product-taste evidence," but the labs carry more weight than that because AI-native evaluators are testing whether you've actually built with the material. The essay proves systematic AI thinking. The labs prove you build with it. Both are required. Pick two: Carrier IQ and UAT Sentinel. Both demonstrate autonomy-threshold judgment, which is the rarest credential at this level. Carrier IQ separates agent assembly from broker evaluation. UAT Sentinel separates automated test execution from human release recommendation. Name those boundaries. The claim: "I've built live agents solo, and in each one I've designed the specific point where agent autonomy ends and human judgment begins."

Misframing risk: presenting all labs with equal weight. Brand Pulse, Retail Velocity, and Creator Scout demonstrate agentic patterns but at lower consequence stakes. Lead with the two that prove the rarest judgment. Mention the others for breadth.

Growth-stage regulated (supporting evidence, moderate confidence). Carrier IQ operates in insurance, a regulated domain. The fact that you built an agent that runs carrier quote flows and keeps the broker in the evaluation seat maps to how regulated companies think about AI adoption. One sentence: "I built a live insurance-intelligence agent where the agent assembles quotes in parallel but the broker makes the call." Then move on.

Misframing risk: walking through multiple labs. Regulated audiences will hear "side projects" and wonder why you're not showing them shipped work with real stakes.

Enterprise platform (supplementary context at best, high confidence). Enterprise evaluators measure organizational impact. Solo-built agents without adoption metrics don't register on that scale. Mention the labs only if asked about AI experience. Frame them as: "Applied research I built to pressure-test agentic interaction patterns before bringing them into a production context." That framing converts "side project" into "R&D discipline."

Misframing risk: showing the labs in a portfolio walkthrough. Enterprise rooms will compare them unfavorably to Alibaba and Thermo Fisher and question your judgment about what to present.

Healthcare/vertical SaaS (supplementary context, moderate confidence). If the healthcare company is not building AI products, the labs don't serve this audience. Skip them entirely. If the company is explicitly building AI into their product, use Carrier IQ as a single reference: "I built a live agent in insurance that keeps the human in the evaluation seat. The pattern translates directly to clinical or regulatory review workflows." One sentence, then move on.

Tier Cards

AI-Native — testing: architectural fluency

1. Essay — threshold credential. Name the five-handoff framework, say "watch, verify, delegate," then move within 90 seconds.
2. Agentic Labs — Carrier IQ + UAT Sentinel. Name autonomy boundaries. Full walkthrough.
3. Thermo Fisher — five-agent vision only. Name the human gate. Skip the supply chain narrative.
4. Alibaba — agentic vision only. Skip the shipped sprint story.

Growth-Stage Regulated — testing: velocity under real constraints

1. Thermo Fisher — full case. Consequence + constraint dissolution + "twelve months, 0-to-1, six partners adopted."
2. Essay — Decision Gate passage only. Name the regulatory human gate.
3. Carrier IQ — one sentence. Regulated-domain AI fluency.
4. Alibaba — one sentence for altitude, then stop.

Enterprise Platform — testing: organizational scale and influence

1. Alibaba — full case. Diagnostic method, organizational influence. Save agentic vision for closing.
2. Thermo Fisher — partner adoption, cross-org complexity.
3. Essay — mention only if AI strategy surfaces. One sentence plus link.
4. Agentic Labs — verbal only if asked. Frame as applied research.

Healthcare / Vertical SaaS — testing: consequence literacy in regulated domains

1. Thermo Fisher — full case. Exception-first design under GDP/cGMP. Earn trust before introducing five-agent vision.
2. Essay — Decision Gate passage only. Bridge from Thermo Fisher to AI thinking.
3. Alibaba — one sentence for altitude.
4. Agentic Labs — Carrier IQ only if the company is building AI products, framed as regulated-domain agent with human-in-the-loop. Otherwise skip.