Every company here scored 7+ on the company rubric, accepted Bay Area candidates in its most recent hiring cycle, and had no qualifying product-design role — Staff+ IC or Director+ leadership — on its public career board as of September 12. Each one is Ignore-until-trigger.
The week after Labor Day is when summer-stalled searches restart and Q4 headcount plans convert into requisitions. Several of these companies closed major rounds in the past four months — Rillet ($100M Series C), Corgi ($106M Series B1), Together AI ($800M Series C), Sierra ($950M) — without posting design leadership. Capital without a design seat is a leading indicator.
This is the third pass. Issue 9 organized dormant companies by design problem; Issue 10 reorganized by trigger type but only re-verified clusters one through three. This issue runs the full integrity check: every board re-scanned, companies with live qualifying roles removed (Harvey, Motive, Scale AI, Synthesia, and eleven others now in the active queue), MaintainX removed after its Autodesk acquisition, Cognition excluded because its $2B financing four days ago is still a hot signal.
How to use this. Set monitoring alerts (LinkedIn, Ashby, Greenhouse) for each company. When a trigger fires, the move type tells you what to lead with, who to find, and how fast to go. Read the tier playbook for that company's archetype before outreach — the tier is listed in each fully developed entry and can be inferred from the cluster for compressed entries. You should not need to start research from zero.
Confidence note. Board-absence checks are high confidence for Ashby, Greenhouse, and Lever inventories that returned successfully. Mercury, Clari, and Transcarent returned empty feeds — their absence finding is moderate confidence. Five companies carry unverified Bay Area eligibility: Weaviate, Cato Networks, Lightricks, SafetyCulture, and Formlabs. Monitor them, but confirm geography before investing outreach time.
1. Clinical Evidence and Care Decisions
AI generates clinical evidence — a diagnosis, a coding recommendation, an authorization decision — and a clinician or care operator must decide whether to trust it before a patient outcome depends on it. Portfolio routing: Red Cross (consequence design, shared records across roles, $847K disbursed) and Thermo Fisher (regulated multi-party operations, 100% adoption), both completed as Product Design Director at BCG DV. For companies where the AI output itself needs grading, lead with an Inference-Aware UX artifact. OpenEvidence's recent EvidenceGrade launch and Suki's Science at Suki initiative show this cluster expanding its design surface without posting design leadership.
-
AKASA. AI for medical coding and revenue-cycle operations — translating clinical documentation into billing codes with human exception review.
- Stage / Size / Capital: Series B/growth; ~300 people.
- Offices: South San Francisco; US remote.
- Value chain: Midstream automation between clinical documentation and payer reimbursement. Design scope concentrates on exception queues and coding-confidence surfaces.
- Best at: Surfacing only the cases that require human judgment, reducing coding backlogs.
- Signals: No recent qualifying signal.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: AKASA's core problem: showing coders exactly why the AI chose a specific code and what evidence supports it, so the coder can confirm or override without re-reading the full chart. Red Cross's role-specific views on a shared record — where caseworkers, supervisors, and finance each see the same disbursement data filtered for their decision — speak to the coder-vs-auditor view split. Juno has designed consequential review workflows where the reviewer is liable for the outcome.
- Verdict: Ignore-until-trigger. Trigger: Staff+ or Director role owning coding evidence or exception review. If trigger fires → Read Healthcare/Regulated playbook. Contact the design or product-function owner within 48 hours; lead with Thermo Fisher (BCG DV) + Output Review artifact.
-
PathAI. AI-assisted pathology connecting model predictions to pathologist review.
- Stage / Size / Capital: Series C/D; ~400 people.
- Offices: Boston; US-remote hiring.
- Value chain: Upstream diagnostic infrastructure — the AI reads tissue; the pathologist confirms.
- Best at: Integrating computational pathology into existing laboratory workflows without disrupting pathologist authority.
- Signals: No recent qualifying signal.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: Pathologists reviewing AI-flagged tissue regions face the same structural challenge as Red Cross caseworkers reviewing disbursement eligibility: high-consequence binary decisions on AI-surfaced evidence, where a false negative costs categorically more than a false positive. Lead with Red Cross's decision-gate architecture and Thermo Fisher's regulated workflow (both BCG DV).
- Verdict: Ignore-until-trigger. Trigger: US-eligible Staff+/Director connecting model evidence to pathologist review. If trigger fires → Read Healthcare/Regulated playbook. Identify the pathology-product design manager; lead with Red Cross decision gates + Thermo Fisher (BCG DV).
-
Clarify Health. Healthcare analytics for clinical, operational, and commercial decisions.
- Stage / Size / Capital: Growth private; ~300 people.
- Offices: San Francisco; US remote.
- Value chain: Downstream analytics — aggregates clinical and claims data into decision-support surfaces for health systems and payers.
- Best at: Turning fragmented healthcare data into actionable operational and clinical intelligence.
- Signals: No recent qualifying signal.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Clarify's challenge is information hierarchy — which metrics surface first, which drill-downs matter, how to prevent data overload in a domain where every number carries clinical or financial consequence. Alibaba's search-result trust signals and homepage information architecture (Head of Design and Research, North America; +20% transactions) are the direct parallel. Juno redesigned expert-user information surfaces at $50B+ GMV scale.
- Verdict: Ignore-until-trigger. Trigger: Staff+ design mandate for evidence-backed care or commercial decisions. If trigger fires → Read Healthcare/Regulated playbook. Route through the senior product leader; lead with Alibaba.
-
ConcertAI. Clinical-research and oncology intelligence connecting real-world data, trials, and treatment evidence.
- Stage / Size / Capital: Growth private; ~400 people.
- Offices: Boston area; US hiring.
- Value chain: Midstream intelligence between clinical data sources and trial/treatment decisions.
- Best at: Connecting real-world oncology evidence to clinical-trial design and treatment selection.
- Signals: No recent qualifying signal.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: ConcertAI's design problem is audit-trail integrity — oncologists and trial sponsors need to trace how evidence was selected, filtered, and weighted before making a treatment or enrollment decision. Red Cross's shared-record architecture across six consolidated systems proves Juno can design for multi-stakeholder auditability under regulatory constraint (BCG DV).
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director role translating model evidence into trial or treatment decisions. If trigger fires → Read Healthcare/Regulated playbook. Contact the clinical-product owner within 48 hours; lead with Red Cross (BCG DV).
-
OpenEvidence. AI clinical decision-support for physicians, with a new EvidenceGrade surface that visualizes evidence strength beneath AI answers.
- Stage / Size / Capital: Series D, $250M at ~$12B valuation; ~400 people.
- Offices: Miami, San Francisco; US remote.
- Value chain: Downstream — the physician-facing layer where AI-generated clinical answers meet clinical judgment.
- Best at: Making the evidence hierarchy beneath an AI-generated clinical answer inspectable and gradable by the physician using it.
- Signals: EvidenceGrade launched July 10 — a direct expansion of the design surface around evidence evaluation. January funding is stale. No qualifying design seat.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 10/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 3. Role (projected) 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: EvidenceGrade is the Trust essay's Output Review pattern made literal — the physician sees the AI's answer, then inspects the evidence quality beneath it before acting. Lead with a specific Inference-Aware UX artifact showing how confidence, source quality, and recency change what the interface presents. Juno has both the conceptual framework (Trust essay's five handoffs) and production proof of consequence design (Red Cross, BCG DV).
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director ownership of evidence grading, citation inspection, or clinician decision surfaces. If trigger fires → Read Healthcare/Regulated playbook. Lead with Inference-Aware UX artifact + Red Cross (BCG DV); identify Head of Product Design.
-
Suki. Ambient clinical intelligence that drafts documentation during care encounters.
- Stage / Size / Capital: Late private; ~300 people.
- Offices: Redwood City.
- Value chain: Downstream — sits in the exam room, listens, drafts. Design scope covers the review-and-correction loop between ambient capture and finalized documentation.
- Best at: Reducing documentation burden without requiring clinicians to change their workflow.
- Signals: Science at Suki launched September 10 — a research initiative for ambient AI evaluation standards. No qualifying design seat.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: Suki's core design tension is correction cost — the clinician reviews an AI-drafted note under time pressure and must decide what to fix, what to accept, and what to flag. The Trust essay's Watch-Verify-Delegate ladder applies, but the specific proof is the staged review architecture Juno designed for Carrier IQ, where an underwriter moves through evidence inspection before committing to a binding quote. Both domains penalize false confidence, and the reviewer cannot defer the decision without creating downstream liability.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director ownership of ambient review, correction, or evidence standards. If trigger fires → Read Healthcare/Regulated playbook. Contact the design leader or clinical-product executive within 48 hours; lead with Watch-Verify patterns + Carrier IQ.
-
Qventus. AI-assisted hospital capacity, scheduling, and discharge coordination.
- Stage / Size / Capital: Growth private; ~300 people.
- Offices: Mountain View; US remote.
- Value chain: Midstream operational layer between hospital data systems and the staff making capacity and discharge decisions.
- Best at: Turning hospital operational data into specific, actionable recommendations for bed management and discharge timing.
- Signals: No recent qualifying signal.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Qventus's design problem is multi-role operational coordination under surge — nurses, bed managers, discharge planners, and physicians all need the same patient-flow state filtered for their decision authority. Red Cross consolidated six legacy systems into one shared record serving field workers, caseworkers, supervisors, finance, and program staff during national disaster response (BCG DV). The parallel is direct.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director mandate for operational assistants or exception escalation. If trigger fires → Read Healthcare/Regulated playbook. Identify the hospital-operations product owner; lead with Red Cross (BCG DV).
-
Cohere Health. AI-assisted prior authorization connecting payers and providers.
- Stage / Size / Capital: Growth private; ~700 people.
- Offices: Boston; US remote (current design expansion is India-based).
- Value chain: Midstream — sits between provider requests and payer decisions, automating the authorization evidence chain.
- Best at: Reducing prior-authorization cycle time by automating clinical-evidence extraction and matching.
- Signals: Design team expanding in India; no US qualifying role. Head of Design is Jon Snydal, leading ~12 product designers.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio value 2.
- Portfolio match: Prior authorization is a two-sided evidence problem — the provider submits clinical justification, the AI evaluates it against payer criteria, and a reviewer decides. Carrier IQ's evidence-backed comparison and approval gate fit this staged-review pattern. Thermo Fisher's regulated multi-party workflow (BCG DV) adds proof of designing across organizational boundaries under compliance constraint. Juno has designed the handoff between automated evidence extraction and human adjudication in domains where both over-approval and under-approval carry distinct regulatory and patient consequences — a specificity most AI design candidates lack.
- Verdict: Ignore-until-trigger. Trigger: Jon Snydal announces a US Staff+ or Director search. If trigger fires → Read Healthcare/Regulated playbook. Contact Jon Snydal within 48 hours; lead with Carrier IQ, supported by Thermo Fisher (BCG DV).
-
Regard. Clinical AI analyzing patient records for diagnoses, documentation, and revenue-integrity opportunities.
- Stage / Size / Capital: Series B/growth; ~300 people.
- Offices: Los Angeles, New York, San Francisco; remote.
- Value chain: Downstream — surfaces diagnostic and documentation opportunities directly to clinicians during care.
- Best at: Identifying missed diagnoses and documentation gaps from existing patient records.
- Signals: No recent qualifying signal.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: Regard's design challenge is alert credibility — the AI surfaces a potential missed diagnosis, and the clinician must evaluate the evidence fast enough to act during the encounter without alert fatigue. Red Cross's consequential decision gates under time pressure are the proof (BCG DV); the Inference-Aware UX artifact showing how confidence and evidence quality change what surfaces is the forward-looking differentiator.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director role owning alert evidence, clinician verification, or correction. If trigger fires → Read Healthcare/Regulated playbook. Identify the design buyer; lead with Red Cross (BCG DV) + Inference-Aware UX artifact.
-
Oshi Health. Virtual multidisciplinary GI care combining clinical services with longitudinal patient software.
- Stage / Size / Capital: Series C; ~300 people.
- Offices: New York; US remote.
- Value chain: Vertically integrated — owns the care delivery and the patient-facing software.
- Best at: Coordinating multiple specialist types (gastroenterologists, dietitians, behavioral health) through a single patient experience.
- Signals: No recent qualifying signal.
- Tier: Healthcare/Regulated.
- Rubric scores: Company 7/12 — AI centrality 1, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 1, Portfolio value 3.
- Portfolio match: Oshi's design problem is longitudinal engagement across multiple care modalities — the patient sees a GI specialist, a dietitian, and a behavioral health provider, and the software must make that feel coordinated. Allē's dual-surface redesign (30M members, 3.2× redemption, BCG DV) solved a version of this: multiple provider types, one consumer, longitudinal engagement mechanics that drive repeat behavior.
- Verdict: Ignore-until-trigger. Trigger: Director+ design mandate or financing-plus-design sequence. If trigger fires → Read Healthcare/Regulated playbook. Contact the care-experience product owner; lead with Allē (BCG DV).
2. Autonomous Security and Threat Response
Companies that detect threats, prioritize risk, and sometimes remediate automatically — faster than a human can inspect every alert. The design question is the delegation boundary from the Trust essay: how far can the system act on its own, and what does the analyst need to see when they intervene? Lead with Brand Pulse (anomaly visibility) and Red Cross (decision pressure, BCG DV) for investigation products. Lead with Carrier IQ (evidence-to-approval) when the system proposes a fix and a human must approve it.
-
Orca Security. Agentless cloud-security platform for asset discovery, attack paths, and risk prioritization.
- Stage / Size / Capital: Series C/growth; ~700 people.
- Offices: Portland, New York, Tel Aviv; US hiring.
- Value chain: Midstream — scans cloud infrastructure and surfaces prioritized risk for security teams.
- Best at: Agentless scanning that discovers assets and attack paths without deploying software on every workload.
- Signals: No recent qualifying signal.
- Tier: Enterprise platform.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Orca's investigation surfaces must show analysts why a specific attack path was prioritized over thousands of others. Brand Pulse's anomaly-visibility architecture — where the system explains why a signal is unusual, not just flags it — addresses this evidence-presentation problem.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director design ownership of AI investigation or remediation controls. If trigger fires → Read Enterprise platform playbook. Identify the security-product design lead; lead with Brand Pulse.
-
Apiiro. Application-security platform using code context to prioritize and remediate risk.
- Stage / Size / Capital: Series B/growth; ~300 people.
- Offices: New York, Tel Aviv; US remote.
- Value chain: Upstream — analyzes code and software supply chain before vulnerabilities reach production.
- Best at: Using code-change context to distinguish real risk from noise in application security.
- Signals: No recent qualifying signal.
- Tier: AI-native.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Apiiro's agent-proposed fixes require developer approval before merging — the same approve/reject/modify pattern as Carrier IQ's quote review states. Juno designed that approval gate for a domain where reversing a bad approval has real financial consequence.
- Verdict: Ignore-until-trigger. Trigger: Staff+ role for agent-proposed fixes or developer approval. If trigger fires → Read AI-native playbook. Contact the application-security product owner within 48 hours; lead with Carrier IQ.
-
Censys. Internet-asset intelligence for discovering and investigating exposed infrastructure.
- Stage / Size / Capital: Series C; ~300 people.
- Offices: Ann Arbor; US remote.
- Value chain: Upstream intelligence — maps the internet's attack surface so security teams can find what they didn't know was exposed.
- Best at: Comprehensive internet scanning that discovers assets organizations don't know they own.
- Signals: No recent qualifying signal.
- Tier: Enterprise platform.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Censys's investigation workflows require entity linkage — connecting a discovered asset to its owner, exposure history, and remediation status. Alibaba's search-result trust signals (+20% transactions) addressed a version of this: helping expert users assess entity credibility from search results before committing to deeper investigation.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director for investigation or remediation workflows. If trigger fires → Read Enterprise platform playbook. Identify VP Product or Head of Design; lead with Alibaba.
-
Obsidian Security. SaaS identity and threat-detection platform for compromised accounts.
- Stage / Size / Capital: Series C/growth; ~300 people.
- Offices: Palo Alto, Newport Beach.
- Value chain: Midstream detection — monitors SaaS identity behavior and surfaces compromised-account evidence.
- Best at: Behavioral detection of identity compromise across SaaS applications.
- Signals: No recent qualifying signal.
- Tier: AI-native.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Identity-risk explanation requires showing analysts what a compromised account did, why the behavior is anomalous, and what the exposure looks like — before the analyst decides to suspend the account. Carrier IQ's evidence-to-approval sequence provides proof for staged evidence presentation before a consequential action.
- Verdict: Ignore-until-trigger. Trigger: Staff+ role owning identity-risk explanation and response. If trigger fires → Read AI-native playbook. Contact the identity-product owner; lead with Carrier IQ.
-
Coalition. Cyber-insurance and active-security platform combining underwriting with continuous risk monitoring.
- Stage / Size / Capital: Series F; ~700 people.
- Offices: San Francisco; US remote.
- Value chain: Vertically integrated — owns both the security posture assessment and the insurance underwriting decision.
- Best at: Connecting real-time security posture to insurance pricing and claims, creating a feedback loop between risk prevention and risk transfer.
- Signals: No recent qualifying signal.
- Tier: Growth-stage platform.
- Rubric scores: Company 9/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 3. Role (projected) 11/15 — Comp 2, Scope 3, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Coalition's unique design problem is continuity between security posture and insurance decision — the same risk data that triggers a security alert also changes an underwriting calculation. Carrier IQ's staged comparison, evidence attachment, and approval gate fit the underwriting-decision surface. Thermo Fisher's regulated multi-party workflow (BCG DV) covers the claims-to-security coordination.
- Verdict: Ignore-until-trigger. Trigger: Staff+ design role for underwriting explanation or incident-to-claim continuity. If trigger fires → Read Growth-stage platform playbook. Identify the insurance-platform design leader; lead with Carrier IQ, supported by Thermo Fisher (BCG DV).
16–20: Corelight (network-detection evidence, Series D, ~300 people, SF), Axonius (cybersecurity asset management, Series E, ~700 people, NY/remote), Cato Networks (cloud-delivered SASE, late private, ~2,000 people, Tel Aviv/Boston — verify Bay eligibility), Huntress (managed cybersecurity for SMBs, growth private, ~700 people, Maryland/remote US), Abnormal Security (behavioral email/identity attack detection, late private, ~700 people, SF/remote).
All five are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board. Portfolio routing:
- Corelight: Lead with Alibaba's expert-user filtering + Brand Pulse anomaly visibility.
- Axonius: Lead with Alibaba's search redesign at $50B+ GMV scale — large-inventory search and trust is the direct match.
- Cato Networks: Verify California eligibility before outreach. Lead with Alibaba enterprise complexity.
- Huntress: The primary user is a Huntress analyst triaging incidents under time pressure — same operational pattern as Red Cross disaster-response triage (BCG DV).
- Abnormal Security: Brand Pulse's anomaly-confidence surfaces are the closest published proof. TinyFish supplies past-tense production context for AI-native credibility but is not portfolio evidence.
3. Evaluation, Inference, and Model-Control Infrastructure
These companies build the tools other companies use to evaluate, observe, serve, and control AI models. The user is a developer or ML engineer. The design problem: making model quality, cost, latency, and failure visible enough that a practitioner can make a release or configuration decision with confidence. Lead with Inference-Aware UX artifacts and Thermo Fisher's enterprise wrapper (BCG DV) for companies selling to enterprise buyers. TinyFish supplies past-tense trace experience but cannot serve as portfolio evidence.
-
Snorkel AI. Data-development and evaluation software for enterprise AI training and quality control.
- Stage / Size / Capital: Series C/D; ~300 people.
- Offices: Redwood City, SF, NY; US remote.
- Value chain: Upstream — the quality of training data and evaluation determines model performance before deployment.
- Best at: Programmatic data labeling and evaluation that replaces manual annotation with scalable, auditable processes.
- Signals: No recent qualifying signal.
- Tier: AI-native.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: Snorkel's hardest design problem is evaluation disagreement — when human annotators and model predictions conflict, the practitioner needs to understand why and decide what to do about it. Lead with a specific Inference-Aware UX artifact showing how disagreement evidence changes a release decision.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director design ownership of evaluation disagreement or dataset quality. If trigger fires → Read AI-native playbook. Contact the product executive; lead with Inference-Aware UX artifact.
-
Labelbox. Training-data, human-evaluation, and model-improvement tooling.
- Stage / Size / Capital: Series D; ~300 people.
- Offices: SF Bay Area; US remote.
- Value chain: Upstream — human evaluation infrastructure feeding model training and improvement loops.
- Best at: Connecting human evaluation workflows to model improvement with provenance tracking.
- Signals: No recent qualifying signal.
- Tier: AI-native.
- Rubric scores: Company 8/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 1. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 3, Portfolio value 1.
- Portfolio match: Labelbox's evaluator queues share the review-state problem Carrier IQ addresses — items move through stages of human review with provenance and re-verification requirements. Juno designed review states where the reviewer's decision has downstream financial consequence.
- Verdict: Ignore-until-trigger. Trigger: Staff+ seat for evaluator queues, provenance, or feedback quality. If trigger fires → Read AI-native playbook. Identify the evaluation-product design owner; lead with Carrier IQ.
-
Braintrust. AI evaluation and observability for prompts, traces, experiments, and production performance.
- Stage / Size / Capital: Series B, $80M; ~100 people.
- Offices: SF, NY, Seattle; remote.
- Value chain: Midstream — sits between model development and production deployment, governing the eval-to-release decision.
- Best at: Making the evaluation-to-release workflow structured and auditable.
- Signals: Current Design Engineer seat does not meet Staff+ threshold. February funding is stale.
- Tier: AI-native.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 3, Design influence 2, Trajectory 1. Role (projected) 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: Braintrust's unit of work is an evaluation or experiment, not an end-user interaction. Lead with a concrete Inference-Aware UX artifact showing how model quality, latency, cost, and regression evidence change a release decision. TinyFish establishes past-tense trace experience but cannot serve as the portfolio artifact.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director product-design mandate over eval-to-release workflow. If trigger fires → Read AI-native playbook. Route to founder/CEO Ankur Goyal; ask who owns product-design evaluation.
24–30: Arize AI (ML/LLM observability, Series C, ~300 people, Berkeley), Pinecone (vector database, Series B, ~300 people, NY/SF), Weaviate (open-source vector database, Series C, ~300 people, Amsterdam — verify US eligibility), Anyscale (Ray-based distributed compute, Series C, ~300 people, SF), Fireworks AI (production inference and serving, Series D, ~300 people, San Mateo/SF), Together AI (open-model training and inference, Series C, $800M at $8.3B, ~300 people, SF), AssemblyAI (speech-to-text APIs, Series C, ~300 people, SF/remote).
All seven are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board, with exceptions noted below. Portfolio routing:
- Arize AI: Trace-inspection workflow is structurally what Juno built at TinyFish — but TinyFish is context, not evidence. Lead with Inference-Aware UX artifact; use TinyFish as past-tense practitioner credibility.
- Together AI: July funding is stale under signal rules, but the $8.3B valuation and thin design org make this a high-priority monitor. Lead with cost/latency/reasoning tradeoff artifacts + Thermo Fisher (BCG DV).
- Fireworks AI: Current Product Designer title is below Staff threshold. Trigger exception: Staff re-leveling, a second design seat, or Director+ posting. Check whether the new role governs cross-product coherence before outreach.
- Pinecone: Lead with Alibaba search-result trust signals for retrieval quality visibility.
- Anyscale: Lead with Inference-Aware UX artifact showing cost-latency-quality tradeoffs.
- AssemblyAI: Lead with Thermo Fisher (BCG DV) for enterprise administration wrapper.
- Weaviate: Verify US payroll eligibility before portfolio work.
4. Developer Agents and Production-Control Surfaces
Developers build, deploy, and debug agents — and the tooling that governs those agents is itself a design surface. Traces, evaluations, rollbacks, permissions, failure recovery. Most infrastructure companies haven't invested in this design layer yet. Lead with Agent Infrastructure as UX artifacts for agent-framework companies; lead with Thermo Fisher (BCG DV) for controlled-release and exception-management products.
-
LangChain. Open-source agent frameworks plus LangSmith observability and deployment tooling.
- Stage / Size / Capital: Series B/growth; ~300 people.
- Offices: NY, SF; distributed US.
- Value chain: Midstream — the framework and observability layer between model APIs and production agent applications.
- Best at: Making agent development iterative through trace inspection, evaluation, and deployment controls.
- Signals: No qualifying role.
- Tier: AI-native.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 3, Design influence 2, Trajectory 1. Role (projected) 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: LangSmith's trace-review and agent-debugging surfaces are exactly the design problem the Agent Infrastructure as UX artifact domain addresses. TinyFish supplies practitioner context; inspectable proof must come from Agentic Labs or a published original artifact.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director product design for trace review, agent debugging, or deployment controls. If trigger fires → Read AI-native playbook. Identify the LangSmith design owner; lead with Agent Infrastructure as UX.
-
Cockroach Labs. Distributed SQL database and managed cloud service built for resilience.
- Stage / Size / Capital: Series F; ~700 people.
- Offices: NY, San Mateo; remote.
- Value chain: Infrastructure — the database layer applications depend on for consistency and availability.
- Best at: Surviving infrastructure failures without data loss or downtime.
- Signals: No qualifying role.
- Tier: Enterprise platform.
- Rubric scores: Company 7/12 — AI centrality 1, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 10/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Cockroach Labs' migration and failure-recovery workflows require exception-first design — the happy path is invisible; the product's value shows up when something goes wrong. Thermo Fisher's exception-first operations (BCG DV) are the direct proof.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director for migration, failure recovery, or AI-assisted administration. If trigger fires → Read Enterprise platform playbook. Contact the database-cloud product leader; lead with Thermo Fisher (BCG DV).
33–40: Starburst (distributed analytics, Series D, ~700 people, Boston/SF), Instabase (document-processing AI, Series C, ~300 people, SF), Workato (enterprise integration/automation, Series E, ~2,000 people, Mountain View), Postman (API and agent-development platform, Series D, ~2,000 people, SF), LaunchDarkly (feature management and release controls, Series D, ~700 people, Oakland), Sentry (application monitoring and debugging, Series F, ~300 people, SF), Supabase (open-source backend platform, Series C, ~300 people, distributed), PostHog (product analytics and experimentation, Series C, ~300 people, remote US).
All eight are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board, with exceptions noted below. Portfolio routing:
- Starburst, Workato: Lead with Alibaba's expert information architecture.
- Instabase: Lead with Red Cross (BCG DV) + Carrier IQ (extraction confidence and review queues).
- Postman: Lead with Agent Infrastructure as UX (Agent Mode and MCP connections).
- LaunchDarkly: Trigger exception: design-leadership appointment following the public departure of a longstanding design/product leader, or Staff+ role for AI release controls. Verify succession before outreach. Lead with Thermo Fisher's controlled-release and exception-first operations (BCG DV).
- Sentry: Lead with trace inspection and reversible-action design.
- Supabase, PostHog: Trigger exception: formal Staff leveling or Director+ design mandate. Current unleveled Product Designer seats don't qualify. Establish level and decision rights before portfolio work.
5. Professional Agents Handling Judgment Work
Agents that draft contracts, research investments, resolve support cases, or manage knowledge — where a human remains accountable for the outcome. The design problem is the delegation boundary: what can the agent do autonomously, what requires human review, and how does the human know when to intervene? Lead with Human-Agent System Design artifacts for agent-configuration products; lead with Thermo Fisher (BCG DV) for multi-party approval workflows.
-
Ironclad. Digital contract lifecycle and contract-intelligence platform.
- Stage / Size / Capital: Series E; ~700 people.
- Offices: San Francisco; remote.
- Value chain: Midstream — sits between legal teams and the contracts they negotiate, review, and execute.
- Best at: Making contract workflows structured and auditable while preserving the flexibility legal teams need.
- Signals: No qualifying role.
- Tier: Enterprise platform.
- Rubric scores: Company 8/12 — AI centrality 2, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 2, AI exposure 2, Portfolio value 3.
- Portfolio match: Contract AI must surface clause provenance and preserve human approval before a legal commitment — the same irreversible-decision architecture as Thermo Fisher's six-partner pharma platform (BCG DV) with its single human QA gate before batch release. Juno designed the approval architecture for a domain where reversing a bad approval has multi-million-dollar consequence.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director seat for AI negotiation, clause provenance, or approval gates. If trigger fires → Read Enterprise platform playbook. Identify the contract-AI design owner; lead with Thermo Fisher (BCG DV).
-
Decagon. AI customer-service agents with configuration, evaluation, and escalation tooling.
- Stage / Size / Capital: Growth private; ~300 people.
- Offices: SF, NY; distributed US.
- Value chain: Downstream — the agent handles customer interactions; the configuration and evaluation layer is the design surface.
- Best at: Making AI agent configuration accessible to non-technical support teams while maintaining quality controls.
- Signals: No qualifying role.
- Tier: AI-native.
- Rubric scores: Company 9/12 — AI centrality 3, Stage/equity 2, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 2, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: Decagon's configuration surface is where a support manager defines what the agent may do, how it escalates, and how its performance is reviewed. The Human-Agent System Design artifact domain addresses this directly.
- Verdict: Ignore-until-trigger. Trigger: Staff+ ownership of agent configuration, performance review, or human escalation. If trigger fires → Read AI-native playbook. Contact the agent-product owner within 48 hours; lead with Human-Agent System Design.
43–50: Cresta (contact-center AI, Series D, ~300 people, Palo Alto — current design hiring is Berlin-only), Rogo (AI research for investment banking, growth private, ~100 people, NY/SF), Notion (collaborative workspace with AI agents, late private, ~2,000 people, SF/NY), Glean (enterprise search and AI agents, Series F, ~2,000 people, Palo Alto), Airtable (configurable database with AI, late private, ~700 people, SF), Palantir (operational AI platform, public, ~3,000 people, Palo Alto), Sierra (enterprise customer-service agents, late private, $950M at $15B+, ~700 people, SF), Pylon (B2B customer-support platform, Series B, ~100 people, SF/NY).
All eight are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board, with exceptions noted below. Key routing:
- Notion: Trigger exception: if Randy Hunt announces a Staff+ search tied to Agents, Workers, or AI-enabled UI, contact within 48 hours. Lead with Human-Agent System Design.
- Glean: Expanded agent product August 26 with agent identity and governance. Trigger exception: if a permanent Head/VP Design is named, start the 30-day executive-hire clock. Interim Head of Design was named ~1 year ago; current status unclear. Resolve the permanent design-leadership structure before personalized outreach.
- Sierra: Route through Head of Product Pol Peiffer while asking who owns design evaluation. Lead with the Trust essay's delegation envelope + Agentic Labs.
- Cresta: Verify US location before outreach.
- Rogo: Lead with Alibaba search + evidence-provenance artifact.
- Airtable: Trigger exception: Staff leveling or Director+ product-design mandate. Current unleveled Product Designer seat doesn't qualify. Confirm title, scope, and trajectory.
- Palantir: Lead with Red Cross decision gates (BCG DV).
- Pylon: Lead with Human-Agent System Design; contact the product founder within 48 hours of trigger.
6. Financial Decisions, Fraud, and Automated Risk
AI makes or recommends financial decisions — underwriting, fraud detection, accounting close, compliance checks — where the decision carries regulatory consequence and the reviewer is personally liable. The design problem is staged evidence: the AI proposes, the human inspects the evidence, and the approval creates a record. Lead with Carrier IQ for every company in this cluster. Pair with Thermo Fisher (BCG DV) when the company also has multi-party regulatory coordination.
-
Rillet. AI-native ERP for accounting close, reporting, reconciliations, and multi-entity finance.
- Stage / Size / Capital: Series C, $100M at $1B valuation; ~100 people. ARR reportedly doubled in three months.
- Offices: San Francisco, New York; remote.
- Value chain: Downstream — the accountant-facing layer where AI-generated reconciliations, journal entries, and close tasks meet human review before financial statements are finalized.
- Best at: Replacing manual accounting-close workflows with AI-generated work products that accountants inspect and approve.
- Signals: August funding is stale under signal rules but recent enough that Q4 design hiring is plausible. Expanding into biotech, healthcare, fintech, logistics, and professional services — each vertical adds regulated accounting contexts the product must represent. No qualifying design seat on the current Ashby board.
- Tier: AI-native.
- Rubric scores: Company 10/12 — AI centrality 3, Stage/equity 3, Design influence 2, Trajectory 2. Role (projected) 12/15 — Comp 2, Scope 3, Craft 3, AI exposure 3, Portfolio value 1.
- Portfolio match: Rillet's hardest design problem is reviewer liability — an accountant inspects an AI-generated reconciliation and must decide whether the evidence is sufficient to approve it before the books close. Carrier IQ's staged comparison, evidence attachment, referral states, and approval gate address this directly: the underwriter reviews AI-generated quotes, inspects the supporting evidence, and commits. Juno designed this for a domain where the approval creates a binding financial record and reversal means regulatory exposure. Thermo Fisher's multi-partner regulated workflow (BCG DV) adds proof of designing across organizational boundaries under compliance constraint.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director product design role on Rillet's primary ATS board. If trigger fires → Read AI-native playbook. Contact the finance-product owner within 48 hours; lead with Carrier IQ, supported by Thermo Fisher (BCG DV).
-
Corgi. Full-stack commercial-insurance platform spanning underwriting, policies, claims, and embedded distribution.
- Stage / Size / Capital: Series B1, $106M at $2.6B valuation, following a $160M Series B in early May; ~100 people.
- Offices: San Francisco, Salt Lake City; distributed.
- Value chain: Vertically integrated — owns underwriting, policy administration, claims, and distribution. Design scope spans the full insurance lifecycle rather than a single workflow.
- Best at: Building a modern commercial-insurance stack from scratch, including trucking and other specialty lines where legacy systems are weakest.
- Signals: Two rounds totaling $266M in May alone. No qualifying product-design role on the current Ashby board; a Staff Brand Designer does not qualify.
- Tier: Growth-stage platform.
- Rubric scores: Company 9/12 — AI centrality 2, Stage/equity 3, Design influence 2, Trajectory 2. Role (projected) 11/15 — Comp 2, Scope 3, Craft 2, AI exposure 2, Portfolio value 2.
- Portfolio match: Corgi's underwriting-to-claims lifecycle requires the same staged-evidence, human-approval architecture as Carrier IQ's quote review — the underwriter inspects AI-generated risk assessments, attaches evidence, and commits to a binding decision. If trucking remains in scope, Cummins/ZED Connect (BCG DV, ~2016) adds directly relevant domain proof: Juno designed a dual-surface fleet management platform covering driver hours, routes, locations, and predictive maintenance for commercial vehicle fleets. Few AI design candidates bring both insurance-decision design and trucking-domain experience.
- Verdict: Ignore-until-trigger. Trigger: Staff+/Director product design role on Corgi's primary ATS board. If trigger fires → Read Growth-stage platform playbook. Identify the insurance-platform design owner; lead with Carrier IQ, supported by Cummins/ZED Connect (BCG DV) if trucking is in scope.
53–60: Highnote (embedded card issuing, Series B, ~100 people, SF), Numeric (AI accounting close, Series B, ~100 people, SF/NY), Oscilar (AI fraud and credit-risk, Series B, ~100 people, Palo Alto), Sardine (fraud and compliance monitoring, Series C, ~300 people, SF/Berkeley), Mercury (startup banking, Series C/growth, ~700 people, SF — board returned no inventory, moderate-confidence absence), Alloy (identity and fraud decision orchestration, Series C, ~300 people, NY), Assured (AI P&C claims, growth private, ~100 people, Palo Alto), Unit (embedded-finance infrastructure, Series C, ~300 people, NY/SF).
All eight are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board. Portfolio routing:
- Numeric, Oscilar, Sardine, Alloy, Assured: Lead with Carrier IQ — each has a staged-evidence, human-approval design problem. Pair with Thermo Fisher (BCG DV) for companies with multi-party regulatory coordination.
- Mercury: Do not act from an aggregator-only listing; confirm a primary board posting first.
- Highnote, Unit: Lead with Carrier IQ + Alibaba enterprise configuration.
7. Enterprise Data, Analytics, and Control Planes
Governing how data flows, who sees what, and whether AI-generated analysis is trustworthy enough to act on. Expert-user products where the design challenge is complexity reduction without information loss. Lead with Alibaba (Head of Design and Research; $50B+ GMV; +20% transactions) for search, filtering, and information-hierarchy problems. Lead with Brand Pulse for anomaly investigation. Lead with Equinox+ (BCG DV) for multi-brand or multi-tenant governance.
61–70: Collibra (data and AI governance, Series G, ~2,000 people, NY/remote), Algolia (search and discovery APIs, Series D, ~700 people, SF/Paris — senior UX leader departed March 2026, successor unclear), Grafana Labs (observability dashboards, Series D, ~2,000 people, NY/remote), Contentful (enterprise content platform, Series F, ~700 people, Berlin/Denver/SF), Monte Carlo (data observability, Series D, ~300 people, SF/NY), commercetools (composable commerce infrastructure, Series C, ~700 people, Munich/Boston), StackAdapt (ML advertising platform, growth private, ~2,000 people, Toronto/distributed US), Cribl (data-engine and observability pipeline, Series E, ~700 people, SF/remote), Elastic (search, observability, and security with agentic investigation products, public, ~3,000 people, SF/Amsterdam), ThoughtSpot (natural-language analytics and AI BI, late private, ~700 people, Mountain View).
All ten are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board, with exceptions noted below. Key routing:
- Algolia: Trigger exception: design-leadership succession search following the March 2026 departure. Verify the successor before choosing an outreach route.
- Elastic: Attack Discovery and agentic SOC expansion create a growing design surface. Lead with Red Cross decision pressure (BCG DV) + Brand Pulse evidence visibility.
- Contentful: Lead with Equinox+'s shared-system/multiple-brand architecture (BCG DV) — five brands, shared tokens and components, distinct brand worlds.
- ThoughtSpot: Lead with Alibaba's expert-user analytics surfaces.
- commercetools: Lead with Alibaba's homepage-search-PDP redesign.
- Grafana Labs, Monte Carlo, Cribl: Lead with Brand Pulse's trace-to-decision pattern for anomaly investigation.
- Collibra: Lead with Alibaba for governance complexity.
- StackAdapt: Lead with Brand Pulse + Retail Velocity.
8. Physical Operations, Autonomous Systems, and Field Exceptions
Autonomous or semi-autonomous physical systems — drones, vehicles, printers, supply chains — where exceptions require human intervention and the cost of a missed exception is physical. Lead with Cummins/ZED Connect (BCG DV, ~2016) as supporting domain proof for IoT, fleet, and connected-operations companies. Lead with Red Cross (BCG DV) for time-critical operational coordination. Lead with Thermo Fisher (BCG DV) for exception-first supply-chain operations.
71–80: Applied Intuition (autonomy simulation and vehicle software, Series F, ~2,000 people, Mountain View — former design leader departed, current reporting unclear), Zipline (autonomous drone delivery, late private, ~2,000 people, South SF), Verkada (connected physical security, late private, ~2,000 people, San Mateo), Tulip (manufacturing application platform, Series C, ~300 people, Somerville/remote), FourKites (supply-chain visibility, Series D, ~700 people, Chicago/remote), Augury (industrial machine-health AI, Series D, ~300 people, NY/distributed), Formlabs (3D printing hardware and fleet software, Series E, ~700 people, Somerville — verify Bay eligibility), Flexport (digital freight forwarding, late private, ~2,000 people, SF), Euclid Power (renewable-energy project management, Series A, ~100 people, NY/remote), project44 (supply-chain visibility and orchestration, Series F, ~700 people, Chicago/distributed).
All ten are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board. Key routing:
- Zipline: Lead with Red Cross (national, time-critical service, BCG DV) then Cummins/ZED Connect (fleet and predictive maintenance, BCG DV).
- Applied Intuition: Current design openings are not formally Staff+. Lead with Cummins/ZED Connect (BCG DV) as supporting proof if a qualifying role appears.
- FourKites, Flexport, project44: Lead with Thermo Fisher's exception-first architecture (BCG DV), supported by Cummins/ZED Connect (BCG DV).
- Tulip: Lead with Thermo Fisher (BCG DV) for manufacturing operations.
- Euclid Power: Founder/product route; lead with Thermo Fisher's six-partner operating model (BCG DV).
- Verkada: Lead with Cummins/ZED Connect (BCG DV) + Carrier IQ.
- Formlabs: Verify geography first.
9. Voice, Generated Media, and Expressive Intent
AI generates something creative, personal, or identity-bearing — voice, video, presentations, documentation — and the user needs control over what it produces without micromanaging every output. The user specifies what they want, not how to get it, and the system interprets. This is where Intent-Based Interaction artifacts matter most. Lead with Equinox+ (BCG DV) for consumer-facing creative products. Lead with Human-Agent System Design for voice agents and conversational avatars.
81–90: ElevenLabs (speech synthesis and voice-agent platform, Series D, ~700 people, London/NY/SF), Cartesia (low-latency voice and multimodal AI, Series A, ~50 people, SF), Pika (consumer AI video generation, Series B, ~100 people, Palo Alto), Tavus (AI video-persona and conversational-avatar platform, Series B, ~100 people, SF), Gamma (AI presentation and document creation, Series B, ~100 people, SF), Deepgram (speech recognition and voice-agent APIs, Series C, $130M at $1.3B, ~300 people, SF/NY), Speechify (text-to-speech and content-listening, growth private, ~300 people, remote/Bay Area), Writer (enterprise generative AI for governed content and agents, late private, ~700 people, SF/NY), Mintlify (developer documentation with AI authoring, Series B, ~100 people, SF/NY), Lightricks (consumer photo/video creation and generative media, Series D, ~700 people, Jerusalem/NY — verify Bay eligibility).
All ten are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board. Key routing:
- ElevenLabs: Lead with Intent-Based Interaction + Human-Agent System Design for voice agents.
- Deepgram: A specific Inference-Aware UX artifact should show real-time latency, turn-detection, and partial-speech states; Equinox+ (BCG DV) supplies consumer-level craft and speed proof.
- Tavus: Lead with Human-Agent System Design + Allē personalization (BCG DV) — identity consent and memory are the hard problems.
- Writer: Lead with Brand Pulse + Output Review artifact for governed content generation.
- Pika, Gamma, Speechify: Lead with Equinox+ (BCG DV) for consumer creative craft and speed.
- Cartesia: Contact the product founder; lead with inference latency and turn-state artifacts.
- Mintlify: Founder route; lead with Agent Infrastructure as UX.
- Lightricks: Verify Bay eligibility before outreach.
10. Cross-Role Coordination and Operating-System Consolidation
Multiple roles — operators, managers, workers, customers, compliance teams — share state, exceptions, and accountability across a single platform. The system must show each role the right information at the right moment without fragmenting the shared record. Lead with Red Cross (BCG DV) for multi-role operational coordination and Thermo Fisher (BCG DV) for multi-partner exception management.
91–100: Tines (enterprise workflow, app, and agent automation with governance, Series C, ~300 people, Dublin/Boston/remote — launched 3B platform July 28), Zania (AI agents for GRC and audit, early growth, ~30 people, Palo Alto), Traba (industrial labor marketplace, Series B, ~300 people, NY/SF), Accord (customer-onboarding and cross-company execution, Series A, ~100 people, SF/Toronto), Automation Anywhere (enterprise RPA and agentic automation, late private, ~2,000 people, San Jose), Clari (revenue forecasting, Series F, ~700 people, Sunnyvale — board returned no inventory, moderate-confidence absence), SafetyCulture (frontline inspections and operational risk, late private, ~700 people, Sydney/Kansas City — verify US geography), NinjaOne (endpoint management and IT remediation, growth private, ~2,000 people, Austin/distributed), MongoDB (document database and Atlas AI platform, public, ~3,000 people, NY/SF), Transcarent (employer healthcare navigation, late private, ~700 people, SF — board returned no inventory, moderate-confidence absence).
All ten are Ignore-until-trigger. Default trigger: Staff+ or Director product design posting on the company's primary ATS board. Key routing:
- Tines: 3B platform expansion creates a growing design surface around governed agent workflows. Lead with Human-Agent System Design + Red Cross (BCG DV).
- Traba: Lead with Red Cross multi-role coordination + Thermo Fisher operations (BCG DV).
- Accord: Founder/product route; lead with Thermo Fisher's six-partner coordination (BCG DV).
- Zania: Founder route; test compensation and function authority early. Lead with Carrier IQ evidence review.
- Automation Anywhere: Lead with Alibaba + TinyFish as past-tense governance context.
- NinjaOne: Lead with Alibaba enterprise complexity + reversible-action design.
- MongoDB: Lead with Alibaba search redesign.
- Clari, Transcarent: Do not act from aggregator-only listings; confirm primary board postings first.
- SafetyCulture: Verify US geography before outreach. Lead with Red Cross + Thermo Fisher (BCG DV).
Priority Monitors
Five companies raised significant capital without posting qualifying design roles: Rillet, Corgi, Together AI, Sierra, and Deepgram. Post-Labor Day hiring plans are being approved now.
Three companies have named design leaders you can monitor directly: Randy Hunt at Notion, Jon Snydal at Cohere Health, and the still-unresolved permanent Head of Design at Glean. When any of them posts a hiring announcement, that's your highest-confidence trigger.
For the other 97: monitor their Ashby, Greenhouse, or Lever boards weekly. The trigger is the posting. The move is in the entry. You should be able to go from notification to outreach in under two hours.
- Cognition's design org: Cognition's $2B financing at a $48B valuation was excluded from this dormant list because the funding signal is still hot — check its career board within the next ten days for a qualifying design seat.
- Glean's permanent design leader: Glean expanded its agent product on August 26 with agent identity and governance policies, but its permanent Head of Design has not been publicly confirmed since an interim appointment roughly a year ago.
- Nooks' active search: Nooks was removed from the dormant list after its founder issued a fresh public hiring announcement alongside a qualifying Head of Design requisition — score it against the full rubric for the active queue.
- MaintainX acquisition impact: MaintainX was removed after public records indicated an August 2026 acquisition by Autodesk, which requires rebuilding its equity-window thesis before any future design role can be scored.