Verification against July 10 audit
Live page checked today against the prior audit. Two findings to verify.
Mandate language: resolved, holding. Browser title still reads "National Disaster Relief Platform · 0→1 Build — Juno Chen." Meta description still frames 0→1 platform design, six systems replaced, national deployment, 1,689 cases. Hero brief rows still surface Product Design Director / GM, 0→1 Build, 6 months, national deployment, $847K disbursed. No regression. The metadata and proof rows remain the hardest-working elements on this page.
Humanitarian framing dominance at scan speed: still open. H1 is still "Designing for the worst day of someone's life." Hero intro still opens with displaced families and disconnected tools before it gets to "I led design and delivery." Nine days, nothing changed. The prior audit's core argument holds: the brief rows cannot override the headline's anchor. They get interpreted through it.
New finding. The prior audit never evaluated the closing section through the AI-adjacent positioning lens. That gap matters more now than it did July 10. The segment of companies hiring design leadership for agent-assisted workflows has grown over the last quarter, and the positioning work across all three teasers has converged on one question: does the page surface the human-system decision architecture these buyers are trying to build? "Complexity lives in the system, not the interface" appears as a pillar kicker in the closing section. After six workflow sections. After a paragraph about data architecture. It is the single strongest signal on this page for that buyer segment, and it sits in the last place anyone will read.
1. Page title
On the page: "National Disaster Relief Platform · 0→1 Build — Juno Chen."
Leads with the artifact, names the mandate type, puts your name last where it belongs. The title does more positioning work than the hero headline. That gap is the entire hero headline problem in one observation.
Recommendation: No change.
2. First screen
On the page: Hero headline reads "Designing for the worst day of someone's life." Hero paragraph opens with displaced families, disconnected tools, and a federal clock before naming you as the lead. Brief rows below carry the mandate data.
This is where the teaser loses its strongest readers. A design VP scanning this page gets the emotional frame first and the leadership frame second. The brief rows are excellent. They sit below a headline that has already told the reader what kind of story this is. The story it tells: someone who cares deeply about humanitarian work. True. Irrelevant at this level. Every candidate at Director+ cares about impact. The differentiator is what you controlled, what you built, and what happened when it shipped under pressure.
The prior audit recommended leading with "6 systems → 1 platform. National deployment in 6 months. Zero retraining at 10× surge." That recommendation stands. The new lens adds something the prior audit missed.
The hero intro names the decision environment: untrained volunteers, surge conditions, federal oversight. Then it stops. The closing pillar section is where the architecture actually becomes visible. Three fields on the intake form trigger eligibility checks, identity verification, and FEMA audit trail generation automatically. The caseworker sees simplicity. The architecture holds everything else. That detail belongs in the hero. Not six sections below it.
That is a human-agent coherence argument before the term existed. Right now it sits in a paragraph that starts with displaced families.
Revised hero headline:
6 systems → 1 platform. Three fields trigger everything.
Revised hero intro, first two sentences:
As Product Design Director / GM, I designed and shipped a national disaster-relief platform that replaced six disconnected legacy systems with one shared record serving three roles that had never shared a system. Three fields on the intake form trigger eligibility checks, identity verification, and FEMA audit trail generation automatically. The caseworker sees simplicity. The architecture holds everything else.
Then the context: federal clock, untrained volunteers, surge conditions. The mission earns emotional weight after the mandate is established.
3. Outcome proof
On the page: The hero Outcome row reads "$847K disbursed · 6→1 system switches · 1,689 cases in 2 weeks." The 10× surge-readiness proof appears only in the closing pillar section, under the heading "10× volume. Zero retraining required."
$847K and 1,689 cases read as program outcomes under the current headline. Under a mandate-forward headline, the same numbers read as design leadership outcomes. The headline determines which interpretation wins.
10× surge-readiness is missing from the hero entirely. The prior audit flagged this. Based on the pattern I've tracked across companies building agent-assisted workflows, this is the outcome most likely to register with AI-adjacent buyers. It proves the system scales without retraining the human operator. That is the exact problem those teams are solving. High-confidence read from hiring-signal patterns, not confirmed by any single buyer conversation.
Revised Outcome row:
$847K disbursed · 6→1 system switches · 1,689 cases in 2 weeks · 10× surge, zero retraining
Four proof points. Each earns its slot.
4. Craft signal
On the page: Six sections organized by user role (Field Volunteer, Caseworker Case Detail, Caseworker Verification, Finance Officer, Supervisor, Program Director). Each section carries an illustration, a quote, and three card captions referencing specific screens. Eighteen screens total across six sections. All six illustration alt attributes are empty.
The role-organized structure is the strongest craft signal on the page. It demonstrates multi-stakeholder platform thinking, which is the actual design leadership skill. Each section shows you understood three completely different jobs well enough to design role-specific views on a shared data model. Platform architecture made visible through the teaser structure itself. Good.
The section quotes are doing the wrong job. "Every added field is a case not opened" is close to a design rationale but stops at philosophy. "I spent more time navigating screens than talking to the family in front of me" is a user pain point. Both are valid inputs to a design process. Neither delivers what a mandate-forward teaser section needs: proof. The constraint, the tradeoff, the consequence. A section quote in this teaser should sound like:
"We cut intake from eight fields to three; eligibility, identity, and audit trail fire automatically. The caseworker never touches verification."
That names what was decided, what was traded, and what the user never has to do. The current quotes describe the problem each section solved. Replace them with the decision and its outcome.
The section headlines do better work:
- "Three fields, not eight" is a decision.
- "Queue, not case-by-case" is a decision.
- "Ledger as audit record" is a decision.
Match the quotes to that altitude.
The empty alt attributes remain unfixed from the prior audit. When images fail to load, or when a hiring manager's browser renders slowly, the alt text is the craft signal. Empty alt says nothing. An alt that reads "Intake form: three visible fields trigger twelve automated checks" restates the design decision in a context where the image cannot. Fill each alt with a one-sentence restatement of the design decision visible in that section's illustration.
Recommendation: Fill alt attributes. Replace section quotes with constraint-tradeoff-outcome statements at the same altitude as the section headlines.
5. Density
On the page: Six role sections, three cards each, eighteen screens. A closing section with three pillar blocks. Two CTAs (hero and closing).
Density is appropriate for a teaser. The six-section structure earns its length because each section represents a different stakeholder in the platform, and the progression from field volunteer to program director tells a scope story a shorter page could not tell.
The problem is sequence. The closing section contains the page's strongest positioning material: "Complexity lives in the system, not the interface," the three-field trigger architecture, the 10× surge proof. All of it appears after the reader has scrolled through six workflow sections. Most scanners will never reach it.
Recommendation: Do not cut sections. Move "Complexity lives in the system, not the interface" into the hero section. Either as a subhead below the revised headline or as the opening line of the hero intro. It is the design-leadership principle that organizes everything below it. Let it do that work from the top.
6. CTA
On the page: "Request the full case study." Appears twice: below the hero intro and at the bottom of the closing section.
Direct, clear, low-friction. Dual placement catches both the fast scanner and the full-page reader.
Recommendation: No change to CTA language or placement. The CTA works. The problem is what the reader believes they are requesting. Under the current humanitarian headline, they are requesting a story about helping disaster victims. Under a mandate-forward headline, they are requesting proof of platform leadership at national scale. Same button. Different click motivation. Fix the headline and the CTA works harder without changing a word.
What matters is what sits behind the click. The full case study needs to deliver the decision architecture, the governance tradeoffs, and the surge-scaling proof that the revised headline promises. The humanitarian stakes become the pressure test for that proof. They are what makes the decisions consequential. They are not the frame.
All six points converge on one positioning problem.
"Complexity lives in the system, not the interface" is a system-design philosophy that says the human operator makes the consequential judgment calls (is this family eligible, does this case need escalation) while the platform handles verification, compliance, fraud detection, and audit trail generation invisibly. Three fields. Everything else automated but governed.
The companies building AI-assisted workflows right now are trying to solve this. Give a human operator the right decision at the right moment without requiring them to understand the system's full complexity. Scale it to 10× volume without retraining. Maintain federal-grade auditability while the human sees only simplicity. These are live problems at every company I track in the agent-assisted workflow space.
You solved them in 2021. Before anyone called it human-agent coherence.
The teaser buries it below six sections of workflow detail and a humanitarian headline. From the hiring patterns I track, an AI-adjacent buyer scanning this page will read a disaster-relief project and move on before discovering that you designed exactly the kind of human-system decision architecture they are trying to build. That is inference from how these buyers scan portfolios, not something the page itself confirms. But the structural evidence is clear: the strongest signal is in the weakest position.
Move it up. Lead with it. The mission context is powerful as proof of stakes. The frame is: you designed a system where untrained humans make consequential decisions under federal oversight, at surge scale, with zero retraining. The complexity lives in the system. The interface is three fields.
That is what this teaser should say at scan speed.
- Amplitude's human-agent workflow: Their Head of Product Design posting explicitly asks the hire to define how agents and humans work together in the same workflows, which is the exact problem the Red Cross teaser's "complexity in the system" principle already answers.
- Vanta's AI-native design org: The Head of Design role asks for agentic UX patterns, quality systems, and launch review frameworks across a ~40-person design organization, making it a buyer that would read the Red Cross surge-scaling proof as directly relevant.
- Gusto names the failure mode: Their Senior Staff Service Designer posting frames AI service design around escalation ownership, AI wrongness, and coherent human-AI-service handoffs, which is vocabulary the revised Red Cross teaser headline should register against immediately.
- EU AI Act Article 14 alignment: The Act's human oversight requirements for high-risk AI systems — understand capacities, detect anomalies, interpret outputs, intervene or stop — map almost exactly to the Red Cross platform's role-specific views on a governed shared record, which strengthens the case for surfacing the design principle earlier.