Regulated and High-Stakes Watchlist
The through-line across all seven: consequences land on a specific person. A clinician's license, an auditor's signature, a security analyst's triage call. The design problem is making the trust decision feel like an actual decision, not a rubber stamp. Two of them, Ambience and Fieldguide, have live postings and company rubric scores high enough to promote into the active queue immediately. Fieldguide's posting is in its hot window right now. The rest are planted flags. Each has a named trigger, and when it fires, you already know which story to tell.
Two promotion flags with deadlines
Fieldguide — Staff Product Designer posted July 10, hot window closes July 31. Company scores 10/12. Act now. Ambience — Staff Product Designer posted June 24, Watch window closes ~August 6. Company scores 11/12. Move before it decays.
1. Ambience Healthcare. Clinical AI platform generating source-grounded documentation, chart intelligence, and real-time clinical decision support across 100+ specialties.
- Stage / Size / Capital: Series C ($243M, co-led by Oak HC/FT and a16z, Jul 2025). OpenAI Startup Fund, Kleiner Perkins, Optum Ventures participating. ~250 employees.
- Offices: San Francisco HQ. SF-based roles confirmed.
- Value chain: Downstream application layer integrating directly into Epic, Oracle Cerner, athenahealth. The design surface is exactly where AI-generated clinical content meets clinician accountability: the moment a provider decides whether to sign the note.
- Best at: Shipping clinical AI at specialty depth. Ambulatory, emergency, inpatient, nursing, each with different documentation requirements and different trust thresholds. The 2026 product cadence tells the story: Chart Chat for Nursing in April, inpatient expansion in May, nursing suite in June. They are adding clinical surfaces faster than most companies can staff design for them.
- Signals: Staff Product Designer posted June 24, 2026 (30 days, Watch window, crosses to Ignore at day 43, approximately August 6). Role owns "end-to-end UX across Clinical AI products." No visible Head, Director, or VP of Design in public profiles. Read that absence carefully. This may be a pre-leadership greenfield. ⚑ Company scores 11/12 with a live Watch-window posting. Promote to active queue. Deadline pressure: act before August 6.
- Tier: Healthcare/Regulated. Tier Playbook: Read Healthcare/Regulated playbook before any outreach.
- Rubric scores: Company 11/12 — AI centrality 3, Stage/equity 3, Design influence 2, Trajectory 3. Role 13/15 — Total comp 2, Scope 2, Craft depth 3, AI exposure 3, Portfolio value 3.
- Portfolio match: Ambience's specific design problem: one AI-generated clinical note serves multiple accountability roles simultaneously. The clinician signs it. The coder bills from it. The compliance officer audits it. Each trusts the same output for different reasons. Lead with Red Cross's disaster-case platform, where one shared record replaced 6 legacy systems and served volunteers, caseworkers, supervisors, and finance officers through different accountability lenses. Specific proof: $847K disbursed, national deployment, full traceability. The design decision that matters here: same underlying data, different trust surfaces per role. A candidate without multi-role trust-surface experience will default to building separate verification tools for each audience. You've shipped the single-record architecture that makes that unnecessary.
- Verdict + action: Ignore-until-trigger in this cluster. ⚑ Promotion flag: company rubric (11/12) and live Watch-window posting qualify this for Watch treatment in the active Signal Watch queue. Trigger for Director+ upgrade: design leadership posting appears, or the absence of visible design leadership is confirmed as greenfield opportunity.
2. Fieldguide. AI-powered audit and assurance platform automating cybersecurity, privacy, and financial audit workflows from planning through review.
- Stage / Size / Capital: Series C ($75M led by Goldman Sachs Growth Equity, Feb 2026). $125M total raised, $700M valuation. ~234 employees.
- Offices: San Francisco HQ. Remote-friendly across NYC, LA, Austin, Seattle, Portland, London, Seoul. US remote roles confirmed.
- Value chain: Midstream platform layer between audit firms and their regulated clients. The trust boundary here is where AI-generated audit work meets human professional accountability. The auditor's license is on the line, and that fact reshapes every design decision about the review surface.
- Best at: Making AI-assisted audit work governable. Goldman's investment thesis explicitly cites "governance, traceability, and clear human oversight built into every engagement." Product architecture language, not marketing copy.
- Signals: Staff Product Designer posted July 10, 2026 (14 days, hot window). Role owns "complex audit workflows including planning, execution, review, and collaboration" and asks for experience "shaping AI-assisted workflows at a systemic level." No Director+ design role posted. ⚑ Live hot posting at a company scoring 10/12. This posting is in its hot window now. Act before day 21 (July 31). By next cycle it crosses into Watch and the advantage starts decaying.
- Tier: Healthcare/Regulated. Tier Playbook: Read Healthcare/Regulated playbook before any outreach.
- Rubric scores: Company 10/12 — AI centrality 2, Stage/equity 3, Design influence 2, Trajectory 3. Role 13/15 — Total comp 2, Scope 2, Craft depth 3, AI exposure 3, Portfolio value 3.
- Portfolio match: Fieldguide's hardest design problem: an auditor reviewing AI-executed work must decide whether to sign off, knowing their professional license backs the output. This is Watch→Verify→Delegate made literal. The AI executes audit tasks. The auditor watches the output, verifies the evidence trail, then delegates trust by signing. Lead with the Trust essay's five-handoff framework and the Watch→Verify→Delegate ladder. Fieldguide's product is the framework enacted in regulated practice. Bridge to Thermo Fisher mySupply for proof you've shipped this pattern: exception-first order management where batch-risk signals appeared before the delivery gate, 100% adoption across 6 pharma partners because the system made accountability feel supported rather than burdensome. No other candidate walks in with a published trust framework that maps to the audit review workflow and regulated adoption proof that it works.
- Verdict + action: Ignore-until-trigger in this cluster. ⚑ Promotion flag: company rubric (10/12) and live hot posting qualify this for Act treatment in the active Signal Watch queue. Window closes July 31. Trigger for Director+ upgrade: design leadership posting, or second design IC posting within 4 weeks confirming org buildout.
3. Abridge. Clinical AI platform generating source-grounded medical documentation using Linked Evidence, connecting AI-produced notes to the specific conversation moments that generated them.
- Stage / Size / Capital: Series E ($300M led by a16z, Jun 2025). Khosla Ventures participating. ~648 LinkedIn profiles. Live at 300+ health systems, 100M+ conversations annually.
- Offices: San Francisco HQ. Pittsburgh, New York City. Hybrid SF roles confirmed.
- Value chain: Downstream application layer integrating into clinical workflows at the point of care. The evidence-to-commitment boundary is the design surface: the moment a clinician decides the AI-generated note is accurate enough to sign.
- Best at: Source grounding. Linked Evidence doesn't just generate a note. It shows the clinician exactly which conversation segment produced each clinical assertion. That is a design problem, not an NLP problem. The June 2026 platform expansion into care delivery, payment, and evidence-based treatment extends the trust surface well beyond documentation.
- Signals: Staff Product Designer posting (Sep 2, 2025) is 326 days old. Dead. But look at what just appeared: Design Operations Manager posted July 21 (3 days) and Product Lead, CDS posted July 22 (2 days). Fresh design-adjacent cluster. Issue #4 flagged Abridge as a watch item for clinical AI trust signal without a qualifying design role. The cluster confirms the design org is actively building. A design IC posting is likely 4–8 weeks out.
- Tier: Healthcare/Regulated. Tier Playbook: Read Healthcare/Regulated playbook before any outreach.
- Rubric scores: Company 10/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 3. Role rubric: scored when qualifying design role surfaces.
- Portfolio match: Abridge's specific design problem is the confidence-grounded verification moment. A clinician sees an AI-generated note with Linked Evidence connecting each assertion to a conversation segment. The design must make the confidence signal legible enough that the clinician can verify in seconds, not minutes, without the speed undermining the quality of the trust decision. Lead with Carrier IQ's broker decision surface, where agent-generated insurance quotes carried confidence scores, source attribution, and comparison data so the human could verify the agent's output before acting. Specific proof: the design made confidence visible at the point of commitment, which is exactly what Linked Evidence needs to do for clinical assertions. Most clinical AI designers optimize for fewer clicks to sign. You've shipped the confidence-calibration surface that makes the sign-off meaningful.
- Verdict + action: Ignore-until-trigger. The design-adjacent cluster (Design Ops Manager + Product Lead) confirms org buildout but doesn't yet include a qualifying design IC or leadership role. Trigger: new Staff/Senior Staff Product Designer posting or Director+ design role. Monitor Abridge's Ashby board weekly. The cluster suggests a design posting is 4–8 weeks out.
4. HackerOne. AI-powered security platform combining bug bounty, penetration testing, and vulnerability disclosure with agentic AI for exploit validation and prompt-injection testing.
- Stage / Size / Capital: Series E ($49M, Jan 2022). ~$160M total raised. 201–500 employees. No recent funding round disclosed.
- Offices: San Francisco HQ. Remote roles confirmed for SF Bay Area, Boston, Seattle, Austin, DC, London.
- Value chain: Vertically integrated. Owns the security researcher marketplace, the testing infrastructure, and the AI validation layer. The triage boundary is where autonomous security findings meet human judgment: a security team deciding whether an AI-flagged vulnerability is real, severe, and worth acting on.
- Best at: Combining human security researchers with AI agents at scale. The 2026 product roadmap is aggressive: Agentic PTaaS (Jan), Agentic Prompt Injection Testing (Mar), H1 Validation (Apr), H1 Platform at AI scale (Jun), Project Glasswing (Jul). Five major AI product launches in seven months. Prompt-injection vulnerabilities on the platform increased 540% YoY.
- Signals: Director, AI Product Design no longer appears on Ashby board. Mirror sources indicate original posting ~June 5, 2026 (49 days, stale past 43-day threshold). Role reported to CPO Nidhi Aggarwal (hired Jun 2025, 13 months in seat). Comp band per mirror: $270K–$300K for SF. Posting pulled. Three possible reads: filled, paused, or being revised. If it reposts, the revised language will tell you what the first search taught the committee.
- Tier: AI-native. Tier Playbook: Read AI-native playbook before any outreach.
- Rubric scores: Company 9/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 3. Role rubric: scored when qualifying role resurfaces.
- Portfolio match: HackerOne's specific design problem is signal-to-noise at speed. A security team reviewing 50 AI-flagged vulnerabilities per day needs to distinguish real exploits from false positives fast enough to act before the window closes, without the false-positive rate eroding trust in the system over time. Lead with Brand Pulse's real-time brand sentiment monitoring surface, where automated signal detection surfaced brand events that humans decided whether to act on. The structure is the same: both products generate a continuous stream of automated findings, and the design problem is making the signal trustworthy enough that humans act on the real ones without being overwhelmed by noise. Bridge to the Trust essay's escalation framework for how trust calibrates over time. Security-domain design candidates rarely bring real-time automated-signal triage design. You have it, and you've published the framework for how trust scales with it.
- Verdict + action: Ignore-until-trigger. The pulled posting is the strongest signal here. It means the search either concluded or the committee learned something. Trigger: role reposted or new Director+ design posting. Monitor HackerOne's Ashby board biweekly.
5. Giga. Real-time AI agent platform for customer experience, deploying voice, chat, and email agents handling complex support and operations workflows for Fortune 500 companies including DoorDash.
- Stage / Size / Capital: Series A ($61M led by Redpoint, Nov 2025). Y Combinator, Nexus Venture Partners participating. ~111 employees.
- Offices: San Francisco HQ. Onsite role confirmed.
- Value chain: Midstream platform layer between enterprise clients and their end customers, operating AI agents across voice, chat, and email. The design challenge concentrates at the escalation boundary: when the agent decides it cannot resolve, must escalate, or must document a compliance-sensitive interaction mid-conversation.
- Best at: Production resolution at regulated scale. 90%+ production accuracy, customers live in under two weeks. Fortune reports live financial-services deployments handling unusual transfers with regulator-required paper trails.
- Signals: Staff Product Designer posted July 15, 2026 (9 days, hot). IC-level at a Series A. Below the Director+ threshold and early-stage on equity window. No design leadership posting. Financial-services vertical exposure creates regulated-workflow relevance.
- Tier: AI-native. Tier Playbook: Read AI-native playbook before any outreach.
- Rubric scores: Company 8/12 — AI centrality 3, Stage/equity 2, Design influence 1, Trajectory 2. Role 12/15 — Total comp 2 (Series A in SF likely clears $200K cash; equity upside uncertain at this stage), Scope 2, Craft depth 2, AI exposure 3, Portfolio value 3.
- Portfolio match: Giga's specific design problem is mid-conversation escalation in regulated contexts. An AI agent handling a financial transfer flags something unusual. Two trust surfaces operate simultaneously: the customer's experience of the handoff, and the institution's compliance record of what the agent decided and why. Lead with Allē's dual consumer/provider surfaces, where the same transaction (treatment redemption) had to work for two fundamentally different trust models. Members saw value and simplicity. Providers saw patient context and urgency sorting. Specific proof: 3.2× redemption rate because the design respected both sides of the trust equation. AI-agent designers typically bring chatbot or conversational UI portfolios. You bring dual-surface trust design where each side of the same interaction carries different accountability requirements.
- Verdict + action: Ignore-until-trigger. Series A execution risk and IC-level role keep this below threshold. Trigger: Director+ design role posted, or Series B announcement with design org language. Monitor Giga's Ashby board monthly.
6. Eltropy. Agentic AI platform for community financial institutions deploying AI agents across 750 credit unions and community banks for lending, collections, fraud resolution, and member engagement.
- Stage / Size / Capital: Funding stage unclear ($25M Series A in 2021; CB Insights lists Series B, data conflict unresolved). ~281 employees.
- Offices: Santa Clara HQ. Director, Design role requires onsite Mon–Thu in Santa Clara. Geographic filter: passes (Bay Area), but 4-day onsite is a lifestyle constraint worth flagging.
- Value chain: Downstream application layer deploying AI agents directly into credit union and community bank member interactions. The member-trust boundary is where an AI agent's action in a lending, collections, or fraud workflow must feel like the institution's judgment, not a bot's.
- Best at: Distribution density in community finance. 750 institutions, 100+ new clients in 2025, and a new fintech Early Access Program letting third parties build agents on the platform. The regulatory surface area is expanding faster than the design org can cover it. Five designers, all in India.
- Signals: Director, Design listed on Greenhouse with $200K–$325K total comp. Posting date could not be verified from the Greenhouse listing. Without a confirmed publication date, the decay window cannot be classified. The role is live but its urgency is unassessable. Company scores 7/12, below the 8/12 threshold. The role rubric below is scored prospectively: it would apply if the company score improves through a funding round that clarifies stage and equity.
- Tier: Healthcare/Regulated. Tier Playbook: Read Healthcare/Regulated playbook before any outreach.
- Rubric scores: Company 7/12 — AI centrality 2, Stage/equity 1, Design influence 2, Trajectory 2. Role 11/15 (prospective) — Total comp 2, Scope 2, Craft depth 2, AI exposure 3, Portfolio value 2.
- Portfolio match: Eltropy's specific design problem: AI agents operating across 750 institutions, each with its own member relationships, compliance posture, and risk tolerance. A collections agent at a rural credit union and a fraud-resolution agent at a mid-size community bank need different trust calibrations, but they run on the same platform. Lead with Alibaba's B2B platform redesign, where trust architecture had to serve 200K+ suppliers with different business contexts across homepage, search, and product detail surfaces. Specific proof: +20% daily transactions, +2.2pt NPS, because the trust design adapted to context rather than imposing a single model. Platform-scale trust that flexes across hundreds of thousands of different business relationships is rare in AI-agent design portfolios. Your Alibaba work is the proof.
- Verdict + action: Ignore-until-trigger. Company rubric (7/12) below threshold. The 4-day onsite requirement and India-based design team raise scope questions worth watching. Funding stage unclear. Trigger: Series C with clear equity terms, or remote/hybrid policy change on the Director role. Monitor Eltropy's Greenhouse board quarterly.
7. Givebutter. All-in-one nonprofit fundraising platform combining donation forms, events, auctions, peer-to-peer fundraising, donor CRM, and PCI-compliant payment processing through Stripe.
- Stage / Size / Capital: Growth-stage private. $50M strategic investment from BVP Forge (Apr 2024). Not a traditional VC series. Equity structure unclear. ~184 employees.
- Offices: Remote-first. San Francisco listed among hub locations.
- Value chain: Vertically integrated. Owns the fundraising surface, the donor CRM, payment processing, and now Wallet and Spend Cards (Jul 2026, powered by Stripe/Celtic Bank). The fund-flow trust boundary is where donation legitimacy, campaign integrity, and post-raise fund movement must be visible to donors, nonprofits, and compliance teams.
- Best at: Consolidating the nonprofit fundraising stack. Most nonprofits stitch together 4–6 tools for donations, CRM, events, and payments. Givebutter ships them as one platform with real-time fraud monitoring and Trust & Safety review. The July 2026 Spend Card launch extends the trust surface from "money in" to "money out."
- Signals: Director, Product Design posted June 15, 2026 (39 days). API shows isListed=true, but the brief flagged a July 20 application deadline. At 39 days the posting sits in its Watch decay window, but the company rubric (6/12) keeps it below threshold regardless.
- Tier: Growth-stage platform.
- Rubric scores: Company 6/12 — AI centrality 1, Stage/equity 1, Design influence 2, Trajectory 2. Role rubric not scored. Company below 8/12 threshold.
- Portfolio match: Givebutter's specific design problem: the Spend Card launch extends the trust surface from donation collection to fund disbursement, and the design must make every dollar traceable from donor intent through nonprofit allocation through spend without adding friction that suppresses giving. Lead with Thermo Fisher mySupply's regulated supply chain, where materials moved through compliance checkpoints from order through delivery and the design made each checkpoint feel like operational support rather than bureaucratic drag. Specific proof: $20M+ annual margin, 100% adoption across 6 pharma partners, 9 sites. Givebutter tracks dollars through compliance checkpoints the way mySupply tracks pharmaceutical materials through regulatory gates. Most nonprofit-tech designers optimize the donation form. You've shipped the end-to-end traceability layer that makes the money accountable after it arrives.
- Verdict + action: Ignore-until-trigger. Company rubric (6/12) well below threshold. Low AI centrality and unclear equity structure are the gaps. Trigger: Series A/B with institutional VC and AI/ML product investment, or the Director role reposts with AI-trust language that would shift the AI centrality score. Monitor Givebutter's Ashby board quarterly.
Things to follow up on...
- Abridge's design-adjacent cluster: Two fresh postings in three days (Design Operations Manager and Product Lead, CDS both published the week of July 21) suggest a qualifying product design role is 4–8 weeks out.
- HackerOne's pulled Director posting: The Director, AI Product Design role disappeared from Ashby after roughly 49 days, and if it reposts with revised language, the revision will reveal what the hiring committee actually learned it needs.
- Fieldguide's Goldman thesis language: The $75M Series C investment memo explicitly names "governance, traceability, and clear human oversight" as product architecture, which is unusually strong investor-level validation of a decision-rights design mandate.
- Giga's regulated financial deployments: Fortune's coverage of the $61M Series A describes live financial-services clients where AI agents handle unusual transfers with regulator-required paper trails, a vertical that could accelerate a design leadership hire faster than the general CX positioning suggests.