Board as of September 19, 2026. 19 verified roles, 15 companies. Six Act, thirteen Watch. Three postings appeared in the last 48 hours: Webflow's Agentic CMS, Handshake's Autonomous Recruiting, Intuit's Mid-Market Studio.
Changes since Issue #11:
- Three companies added: EvenUp (legal AI, Series E, Act), Unify (GTM agents, Series B, Act on rubric despite early stage), Cherry (patient financing, Watch).
- Affirm's agentic-experiences posting is dead. The Greenhouse URL redirects to an error page; the role was 68 days old before it closed.
- Instrumentl returns from Issue #10. Posting is live and now explicitly requires agentic-surface experience.
The organizing pattern: nearly every Act-qualifying role posted in the last three weeks answers the same question differently — who decides when agent work is good enough to trust? Clusters sort by that authority variant. Highest Act density first.
Agent-Mediated Human Decisions
Both roles are at Handshake, sitting on opposite sides of the same system. In Autonomous Recruiting the agent acts on behalf of employers. In Student Experience the agent advises and the student decides. Both Act.
1. Handshake. AI career platform — autonomous agents matching students with employers. Series F, $3.5B valuation, ~$430M raised, 500–1,000 employees. SF hybrid.
- Value chain: Two-sided marketplace where the matching algorithm is the product.
- Best at: The dataset — 14M+ students, 1,500+ universities — that makes autonomous recruiting viable.
- Signals: Staff PD, Autonomous Recruiting — 2 days ★ NEW. Four concurrent SF design postings. Revenue tracking toward high hundreds of millions (TechCrunch); the "tripled ARR" claim is employer-reported and unverified at baseline.
- Why now: Four design postings in two weeks alongside accelerating revenue reads as org expansion (inferred from posting cluster, not confirmed). The Autonomous Recruiting role is 2 days old, which means you are ahead of the applicant volume rather than behind it.
- Tier: AI-native — read AI-native playbook before outreach.
- Scores: C11 (AI 3, Stage 3, Influence 2, Trajectory 3) · R14 (Comp 3, Scope 3, Craft 3, AI 3, Portfolio 2).
- Portfolio match: Allē's dual-surface architecture. The provider recommends, the consumer decides, and each surface needs different trust signals drawn from the same underlying system. Handshake's recruiting agent works the same way: the employer side optimizes for reach, the student side has to preserve the student's authority over their own career decision. Juno built that split at 30M-member scale with 3.2× redemption lift, which shows both sides of a recommendation system holding without one eating the other's credibility. Candidates with agent-flow experience usually have it on a single surface.
- Act. Route to Battlecard. The Issue #11 Opportunity Brief develops the Student Experience positioning; Autonomous Recruiting is the higher-stakes sibling.
2. Handshake. Company details: see entry 1. Staff PD, Student Experience — 15 days.
- Scope: Agent shapes career decisions for students. Recommendations, not actions.
- Scores: C11 · R14 (Comp 3, Scope 3, Craft 3, AI 3, Portfolio 2).
- Portfolio match: Brand Pulse's evidence-beside-synthesis pattern — hourly scores with every source visible next to the agent's written narrative. A student weighing career recommendations needs the same pairing: the suggestion plus the trail that produced it. Brand Pulse is the one piece in the portfolio where the human is consuming an agent's judgment about their own situation rather than about a task.
- Act. Different authority surface from Autonomous Recruiting. Lead with whichever opens first.
Agent-Generated Content Governance
The agent produces something visible — a web page, a UI component — and someone has to decide whether it ships. Webflow governs full-page output. Cherry governs the component system that limits what the agent can produce in the first place.
3. Webflow. Visual web development platform adding agentic site-building. Series C, $4B valuation (2022), ~$335M raised, 900+ employees. Distributed, 30+ designers.
- Value chain: Midstream platform between design intent and published web output. AI moves the value from manual building to output governance.
- Best at: Professional web development without code. The Agentic CMS bet extends that to AI-generated output.
- Signals: Staff PD, Agentic CMS — 1 day ★ NEW. CPO and SVP AI Transformation appointed July 2026. Reports to Senior Director of Product Design; Kevin Wong is the plausible evaluator (inferred from title, not confirmed).
- Why now: 1 day old, and it is the first public design hire visible since the July leadership change took effect. Two months is about how long a new CPO takes to name their first headcount.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C10 (AI 2, Stage 2, Influence 3, Trajectory 3) · R14 (Comp 3, Scope 3, Craft 3, AI 3, Portfolio 2).
- Portfolio match: The Alibaba.com PDP redesign, specifically the commitment moment where a buyer decides to place a cross-border order worth tens of thousands. Webflow's Agentic CMS sits at the same inflection: a user deciding whether to publish output the system generated. The +20% transaction lift is measured evidence at the point of irreversible action, which is a different claim than having designed a review flow.
- Act. 1 day old. Route to Battlecard immediately.
4. Cherry. Patient financing platform for healthcare providers — BNPL for dental, medical, veterinary. Series C, $2B+ valuation, 500–1,000 employees. Remote US.
- Value chain: Downstream fintech application in healthcare payments. The design-systems role governs agent output across five product surfaces.
- Best at: Vertical BNPL in healthcare, with provider and consumer surfaces in a regulated payment context.
- Signals: Staff PD, AI Design Systems — 15 days. Founding design-systems seat. Claude-first: agents produce on-system UI output. No published compensation.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C7 (AI 1, Stage 2, Influence 2, Trajectory 2) · R11 (Comp 2, Scope 2, Craft 3, AI 2, Portfolio 2).
- Portfolio match: Carrier IQ, where the system produces quotes inside defined parameters and surfaces the cases that fall outside them. At Cherry the design system is the constraint layer on Claude-generated UI, which makes this a compliance mechanism wearing the clothes of a component library.
- Watch. Company score (7) below Act threshold, and AI centrality is genuinely low here. Monitor for a design-leadership posting.
Agent Execution in Consequential Workflows
The agent takes action where mistakes cost money, create legal exposure, or break compliance. The design problem is the approval threshold: at what confidence level does the agent stop and ask a human? Zip and EvenUp are Act; ServiceNow and Pulley are Watch.
5. Zip. Procurement orchestration platform — intake, approvals, vendor management, and now AI Superagents for contract review, invoice coding, and vendor negotiation. Series D, $2.2B valuation (Oct 2024), $371M raised, 500–1,000 employees. SF.
- Value chain: Midstream orchestration between budget holders and vendors. Agent actions run through policies, approval paths, and audit trails.
- Best at: Governed agent execution in procurement. VentureBeat confirmed the agents operate inside existing compliance frameworks rather than around them.
- Signals: Staff PD, AI Agents & Chat — 11 days. The June 2026 Superagent launch is the product moment behind it.
- Why now: Launch in June, design role in September. That sequence usually means the governance surface shipped rough and someone senior has since argued successfully for a designer on it.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C11 (AI 3, Stage 3, Influence 2, Trajectory 3) · R15 (Comp 3, Scope 3, Craft 3, AI 3, Portfolio 3).
- Portfolio match: Carrier IQ's approval gates — the point where automated quote processing pauses for human verification before it commits. Zip runs that pattern at higher dollar values across contract terms, spend thresholds, and vendor selection, and the audit-trail requirement maps directly. Thermo Fisher's $20M+ margin outcome adds the evidence that she designs for financial consequence, not just financial data.
- Act. 11 days old, well inside the window. Route to Battlecard.
6. EvenUp. AI platform for personal-injury law firms — intake, document drafting, workflow automation, case resolution. Series E, $2B+ valuation (Oct 2025), $385M raised, 500–1,000 employees. SF hybrid.
- Value chain: Vertically integrated. The AI generates the legal documents that are the product, so design scope covers intake through filing.
- Best at: Making AI-drafted legal documents reliable enough that attorneys will put their name on them.
- Signals: Director, Product Design — 9 days ★ BOARD ADDITION. Reports to Head of Design. Roughly 50% management, 50% hands-on strategic design. $202–238K plus equity.
- Why now: 9 days old. Series E a year ago, 500+ headcount, and the design team is adding its first Director under an existing Head of Design. That structure suggests a function large enough to need a layer, which is the moment when scope is still negotiable.
- Tier: AI-native — read AI-native playbook before outreach.
- Scores: C11 (AI 3, Stage 3, Influence 2, Trajectory 3) · R13 (Comp 2, Scope 2, Craft 3, AI 3, Portfolio 3).
- Portfolio match: Thermo Fisher's exception-first design. 197 orders, surface the 3 that matter. EvenUp drafts hundreds of documents where most are fine and the exceptions create liability, so the interface problem is the same: high-volume automated output that has to route the right small subset to a human. The $20M margin came out of a regulated domain, and 100% partner adoption is the part that matters — the exception architecture earned trust from people who had reason to withhold it.
- Act. 9 days old. New to the board. Route to Battlecard.
7. ServiceNow. Enterprise workflow platform — IT, security, HR, CRM, customer service. Public, 20,000+ employees. Santa Clara.
- Value chain: Midstream enterprise platform orchestrating workflow across departments. AI agents automate inside existing governance structures.
- Best at: Enterprise workflow automation at scale with compliance built into the platform layer.
- Signals: Four design postings in 17 days. Director PD (ITSM, 13 days), Senior Staff PD (Security/Risk, 17 days), Director PD (Customer & Partner Experience, 16 days), Principal PD (SalesCRM, 16 days).
- Tier: Enterprise platform — read Enterprise playbook before outreach.
- Scores: C8 (AI 2, Stage 2, Influence 2, Trajectory 2). ITSM Director: R14 (Comp 3, Scope 3, Craft 3, AI 3, Portfolio 2). Security/Risk: R13. CX: R13. SalesCRM: R12.
- Portfolio match (ITSM Director, the strongest of the four): the Red Cross platform's multi-role architecture, where three roles shared one record and the system absorbed 10x volume with no retraining. ITSM incident management has the same shape. 1,689 cases in two weeks is operational velocity under genuine pressure, not a pilot.
- Watch. Company score (8) below Act. The cluster is the real signal: four coordinated postings often precede a VP or Head of Design search within a quarter. Monitor for that consolidation.
8. Pulley. Equity management platform for startups — cap tables, 409A valuations, compliance. Growth-stage, SF.
- Signals: Senior/Staff PD — 39 days. Extended — justify: specialized equity-management domain with a thin design candidate pool. Verify still active before outreach.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C8 (AI 2, Stage 2, Influence 2, Trajectory 2) · R11 (Comp 2, Scope 2, Craft 3, AI 2, Portfolio 2).
- Portfolio match: Thermo Fisher's multi-stakeholder platform. Founders, investors, and employees looking at the same equity data with different trust requirements is structurally the same problem as six pharma partners seeing shared performance data for the first time.
- Watch. 39 days is the concern. Do not spend outreach time until you confirm the posting is still active.
Agent-Assisted Knowledge Work
The largest cluster. The agent assists a knowledge worker who keeps primary authority over the outcome, so the design question becomes one of visibility: can the human see enough of what the agent did to trust the result? Unify is the only Act here; the other five are Watch.
9. Unify. AI-powered GTM platform where agents and sellers work together across prospecting, prioritization, personalization, and follow-through. Series B, $260M valuation (Jul 2025), ~$59M raised, ~150 employees. SF or NYC hybrid.
- Value chain: Midstream GTM tooling between CRM data and outbound execution. The agent does prospecting work; the seller keeps judgment.
- Best at: Designing for agent-human collaboration rather than agent replacement. The posting states this outright, which is unusual.
- Signals: Staff PD — 19 days ★ BOARD ADDITION. $230–280K plus equity. 8x revenue growth in the year before the Series B.
- Why now: 19 days puts this two days from the Watch boundary. The comp band is posted and high for the stage, which tends to mean a short, competitive search.
- Tier: AI-native — read AI-native playbook before outreach.
- Scores: C10 (AI 3, Stage 1, Influence 3, Trajectory 3) · R15 (Comp 3, Scope 3, Craft 3, AI 3, Portfolio 3).
- Portfolio match: Brand Pulse, where the agent monitors and synthesizes continuously while the human decides what to act on. That is the daily texture of a seller using Unify: agent work visible, steerable, interruptible. Brand Pulse is the only piece where Juno designed for continuous watching rather than discrete review, and the Trust essay's Watch-Verify-Delegate ladder describes the progression sellers move through as they learn how far to let the prospecting agent run.
- Act. Series B is early and the stage score reflects that, but the rubric clears everywhere else. Move this week.
10. Slack. Workplace messaging and collaboration (Salesforce). Public, SF.
- Signals: Principal Design Architect — 16 days. The Issue #11 Battlecard works through the multi-party authority problem in detail.
- Tier: Enterprise platform — read Enterprise playbook before outreach.
- Scores: C8 (AI 2, Stage 2, Influence 2, Trajectory 2) · R15 (Comp 3, Scope 3, Craft 3, AI 3, Portfolio 3).
- Portfolio match: Brand Pulse's parallel-agent architecture, with multiple sources monitored at once and scored as data lands. Slack's multi-agent workspace needs the same thing on a shared surface, where several humans evaluate several agents' output at the same time. The Battlecard maps the specifics.
- Watch. Highest role score on the board at R15, but the company score of 8 holds it at Watch. Prep is already written if this converts.
11. Stripe. Payment infrastructure. Pre-IPO, SF.
- Signals: Staff PD, Data & AI — 36 days. Extended — justify: named target, pre-IPO with significant equity, specialized team. Verify still active.
- Tier: Enterprise platform — read Enterprise playbook before outreach.
- Scores: C8 (AI 2, Stage 2, Influence 2, Trajectory 2) · R14 (Comp 3, Scope 3, Craft 3, AI 2, Portfolio 3).
- Portfolio match: Carrier IQ again, for the authorization layer over automated financial processing — parameters, then a pause for human verification before commitment. Stripe's Data & AI team builds where accuracy is not negotiable, and the Trust essay supplies the conceptual language for how trust behaves when agents touch sensitive financial data. Alibaba.com carries the enterprise-scale half.
- Watch. 36 days. Confirm the posting is live before you spend anything on it.
12. Intuit. Financial software platform — QuickBooks, TurboTax, Mailchimp. Public, Mountain View.
- Signals: Principal PD, Mid-Market Studio — 2 days ★ NEW.
- Tier: Enterprise platform — read Enterprise playbook before outreach.
- Scores: C8 (AI 2, Stage 2, Influence 2, Trajectory 2) · R13 (Comp 3, Scope 2, Craft 3, AI 2, Portfolio 3).
- Portfolio match: Alibaba.com's enterprise trust architecture, where B2B buyers commit money on the strength of platform-generated information. Mid-market maps to Alibaba's SME buyer segment far more precisely than Intuit's consumer products do, and +20% transactions is the number that speaks to that evaluator.
- Watch. Fresh posting with unclear scope. Watch for the scope to firm up.
13. Instrumentl. Grant discovery and management platform for nonprofits, now with agentic surfaces. $55M growth investment from Summit Partners (Apr 2025), profitable, 5,500+ customers, 51–200 employees. Remote US/Canada.
- Signals: Staff PD — 17 days. $200–230K plus equity. The posting now requires agentic-surface experience, which it did not when Issue #10 first surfaced it. A rewritten requirement usually means the first pass brought in the wrong pool.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C8 (AI 2, Stage 2, Influence 2, Trajectory 2) · R11 (Comp 2, Scope 2, Craft 3, AI 2, Portfolio 2).
- Portfolio match: Red Cross. Nonprofit domain, funding decisions that matter to the people waiting on them, six systems consolidated into one. $847K disbursed across 1,689 cases in a national deployment is sector-specific proof that almost no other candidate for this role will have.
- Watch. Profitable with a growth investor is a different risk profile from venture-backed: less equity upside, more floor.
Agent as Infrastructure User
One role, and it belongs in its own cluster. Here the agent is the user, consuming APIs on behalf of humans, and the design question becomes access governance: what should the agent be allowed to reach?
14. Apollo GraphQL. API platform — GraphQL infrastructure for building, managing, and governing APIs. Series D, $1.5B+ valuation (Aug 2021), $183M raised, ~220 employees. Remote US.
- Value chain: Upstream infrastructure. Apollo is the governance layer between data sources and consumers, and as agents become the dominant consumers, that layer becomes the trust boundary.
- Best at: API governance at the federation layer. Agent-as-API-consumer is the company's growth thesis.
- Signals: Senior Staff PD — 9 days. $200–240K plus equity. Covers developer experience, API governance, and controls over what agents can reach.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C9 (AI 2, Stage 2, Influence 2, Trajectory 3) · R14 (Comp 2, Scope 3, Craft 3, AI 3, Portfolio 3).
- Portfolio match: the Trust essay's five handoffs, specifically the handoff where a system decides how much authority to grant an automated actor. The agent-as-user problem is new enough that most candidates will arrive without any framework for it, and Juno has both the framework and shipped governance UI for automated actors in a constrained domain (Carrier IQ).
- Watch. Company score of 9 sits one point under Act, and the 2021 Series D is aging in a way that affects the equity math. The role itself is the most unusual design problem on this board and would produce a portfolio piece nobody else has. If the posting holds past 21 days, build the Battlecard.
Outside the Authority Pattern
Two roles where AI is present but the core design challenge belongs to the product domain rather than the authority question. They are here so the edge of the pattern is visible.
15. Gusto. Payroll, benefits, and HR platform for SMBs. Growth-stage, SF.
- Signals: Senior Product Design Manager, Banking — 15 days.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C8 (AI 1, Stage 2, Influence 2, Trajectory 3) · R13 (Comp 3, Scope 2, Craft 3, AI 2, Portfolio 3).
- Portfolio match: Thermo Fisher's regulated 0-to-1 build. Banking product design inside an established platform needs compliance-first thinking and multi-party approval flows, and the $20M margin outcome in pharma is the closest analogue in the portfolio.
- Watch. AI centrality of 1 is the ceiling here. Monitor for AI-specific product expansion.
16. Airwallex. Global payments infrastructure — cross-border payments, treasury, spend management. Growth-stage, Melbourne HQ with global presence.
- Signals: Head of Product Design — 18 days. Flag: confirm SF/US eligibility before outreach. Melbourne-headquartered, so geographic mismatch is possible.
- Tier: Growth-stage platform — read playbook before outreach.
- Scores: C9 (AI 1, Stage 2, Influence 3, Trajectory 3) · R10 (Comp 2, Scope 2, Craft 2, AI 1, Portfolio 3).
- Portfolio match: Alibaba.com at $50B+ GMV, same cross-border complexity and the same trust demands on international transactions. +20% transactions and +2.2pt NPS are the numbers that travel here.
- Watch. Lowest role score on the board at R10. Nothing goes out until geographic eligibility is confirmed.
Things to follow up on...
- Slack's September product release: The Dreamforce announcement added channel-based agent participation, voice actions, and human sign-off for high-stakes code deployment — all directly relevant to the Principal Architect role's authority-design mandate.
- Webflow Source launch context: Forbes coverage frames Source as Webflow's shift from page-building toward governed agent workflows for marketing teams, which gives the Agentic CMS outreach a product-strategy hook beyond the job description.
- Zip's governed-agent architecture: VentureBeat reporting confirms Zip's agents operate within existing compliance frameworks for contract review, invoice coding, and vendor negotiation — useful for grounding the Battlecard in the company's own governance language.
- ServiceNow's Agent Studio release: The September reimagined Agent Studio introduces visual workflow mapping and side-by-side build-and-test, which may signal what the ITSM Director role will actually own in its first quarter.