Outreach Target List

Candidates across four channels, verified against current public sources

Strict-core ICP (50 to 500 engineers, SaaS, US, Docker/K8s, ships agents) plus adjacent bands. Most companies that adopted Anthropic Managed Agents, Devin, Composio, or LangSmith are buyers, not lost-prospects. They bought a slice of the problem, not the org-wide substrate. Vendor neutrality is the v4 pitch. Lost-prospect is reserved for direct vertical competitors, public self-builders, and competitor customer pools. Status and notes are tracked per company in companies.ts; edits land via commit.

Total
159 / 159
Contacted
0
Scheduled
0
Done
0
Passed
0
Channel:
Layer:
Status:
CompanyChannelLayerSizeWhy this candidateStatusNotesFind buyer
ChamberYC/FundedPractitioner<50YC W26. GPU optimization via agentic AI; monitors clusters, predicts resource needs, reallocates GPUs in real time. Peer perspective on agent runtime. source ↗NewLinkedIn ↗
ArmatureYC/FundedPractitioner<50YC W26. Automated agent testing across MCPs/CLIs and major harnesses. Direct adjacency — their customers are agent operators. source ↗NewLinkedIn ↗
CascadeYC/FundedLost-prospect<50YC W26. Agent safety/reliability; distills proprietary intelligence for autonomous systems. Direct competitor adjacency — find their customers via case studies. source ↗NewLinkedIn ↗
SalusYC/FundedLost-prospect<50YC W26. Guardrails to validate agent actions — overlaps with v4's approval/policy framing. Direct competitor; their customers are prospects. source ↗NewLinkedIn ↗
SentrialYC/FundedPractitioner<50YC W26. Agent reliability monitoring. Peer perspective; sees the same compliance/audit pressure. source ↗NewLinkedIn ↗
Terminal UseYC/FundedPractitioner<50YC W26. 'Vercel for background agents.' Direct runtime adjacency. source ↗NewLinkedIn ↗
TensolYC/FundedPractitioner<50YC W26. Multi-agent orchestration. Peer in the runtime category. source ↗NewLinkedIn ↗
ashrYC/FundedPractitioner<50YC W26. Automated multi-modal agent testing. Adjacent. source ↗NewLinkedIn ↗
CanaryYC/FundedPractitioner<50YC W26. AI QA engineer that understands codebases. Adjacent — practitioner-rich audience. source ↗NewLinkedIn ↗
EmDashYC/FundedPractitioner<50YC W26. Open-source agentic dev environment. OSS audience, runtime-aware. source ↗NewLinkedIn ↗
MendralYC/FundedBuyer<50YC W26. AI agents for CI failure diagnosis. Agents acting on consequential systems by definition; small enough for design-partner conversation. source ↗NewLinkedIn ↗
Ressl AIYC/FundedBuyer<50YC W26. AI agents for Salesforce configuration. SaaS customer base where agents act on consequential systems. source ↗NewLinkedIn ↗
FenrockYC/FundedBuyer<50YC W26. Financial-crime compliance agents. Their compliance posture aligns tightly with v4's audit/approval framing. source ↗NewLinkedIn ↗
IncidentFoxYC/FundedBuyer<50YC W26. AI-driven incident management; agents acting on consequential systems by definition. source ↗NewLinkedIn ↗
ArclineYC/FundedBuyer<50YC W26. Legal-document drafting agents. Compliance-adjacent buyer. source ↗NewLinkedIn ↗
SpongeYC/FundedPractitioner<50YC W26. Financial infrastructure for the agent economy. Peer perspective. source ↗NewLinkedIn ↗
OrthogonalYC/FundedPractitioner<50YC W26. Agentic payments for APIs. Adjacent to credential mediation conceptually. source ↗NewLinkedIn ↗
Catena LabsYC/FundedBuyer<50$30M Series A May 20, 2026 (Acrew + a16z crypto). AI-native bank; humans set spending limits / recipients / caps for AI surrogates. Aligns conceptually with approval gates. source ↗NewLinkedIn ↗
AlterYC/FundedLost-prospect<50YC S25. Zero-trust identity + RBAC/ABAC for AI agents with policy enforcement and audit trails. Closest direct competitor surfaced in research; founders ex-Goldman / ComputeAI. source ↗NewLinkedIn ↗
Cursor (Anysphere)YC/FundedBuyer100-300Pure AI-coding company; agents in product. Their own platform team has felt the credential/audit shape firsthand. NewLinkedIn ↗
Codeium / WindsurfYC/FundedBuyer100-300Coding agent at scale. Enterprise customers force them to think about credential boundaries on customer code. NewLinkedIn ↗
SourcegraphYC/FundedBuyer200-400Cody agent in enterprise contexts. Enterprise-readiness pain is well-documented in their writing. NewLinkedIn ↗
SierraYC/FundedLost-prospect300-600$15.8B valuation, $150M ARR (Feb 2026), $950M Series C (May 2026). Claims '40% of Fortune 50' as customers. Won't buy; will have the strongest opinion on what an exec layer should look like. source ↗NewLinkedIn ↗
DecagonYC/FundedLost-prospect200-400$4.5B valuation (Jan 2026). AI support agents at enterprise scale. Built their own substrate; lost-prospect with strong opinion data. source ↗NewLinkedIn ↗
CrestaYC/FundedLost-prospect200-400AI for contact centers; serious compliance posture. Built their own platform; lost-prospect. source ↗NewLinkedIn ↗
Hex TechnologiesYC/FundedBuyer100-200Hex Magic agent runs SQL against customer data. Credential/audit story is load-bearing for sales. NewLinkedIn ↗
GleanYC/FundedBuyer300-600Enterprise search + assistant. Mid-market sales motion means SSO/audit are table stakes; agent expansion creates the C3 trigger. NewLinkedIn ↗
Modal LabsYC/FundedPractitioner50-150Compute substrate for AI workloads. They see the credential leakage their customers worry about. Peer-aware audience. NewLinkedIn ↗
BasetenYC/FundedPractitioner50-100ML deployment platform; their customers run inference for agents. Peer perspective. NewLinkedIn ↗
Fireworks AIYC/FundedPractitioner50-150Inference provider; their customers are agent operators. Adjacent to the buyer story. NewLinkedIn ↗
Together AIYC/FundedPractitioner100-200Inference provider; sees agent traffic shape. Adjacent. NewLinkedIn ↗
AnyscaleYC/FundedBuyer200-400Ray-based ML infra. Customer base overlaps with agent-running orgs. NewLinkedIn ↗
LangChainYC/FundedLost-prospect50-150LangSmith ships agent observability and traces. Lost-prospect: their customers chose LangSmith for observability — ask why audit is or isn't separable. NewLinkedIn ↗
LlamaIndexYC/FundedLost-prospect<50[adj] AI-native small. Building agent infra commercially. NewLinkedIn ↗
MintlifyYC/FundedBuyer100-200$45M Series B (2026). Docs platform with AI-readable infrastructure; thesis that 'being findable by AI is a budget line item.' Peer-infra, design-partner-friendly. source ↗NewLinkedIn ↗
VercelYC/FundedBuyer500-1000[adj] Larger. Ships v0 + AI SDK; agents in their own ops. Strong design-partner culture. NewLinkedIn ↗
ReplitYC/FundedBuyer200-400Replit Agent in product. Agents acting on consequential systems by construction; users connect real services. NewLinkedIn ↗
TesslYC/FundedBuyer<50[adj] Mik Kersten's AI-native dev co. Strong opinions on agent substrate. NewLinkedIn ↗
BrowserbaseYC/FundedPractitioner<50[adj] Browser infrastructure for agents. Sees the credential exposure problem directly. NewLinkedIn ↗
AnonYC/FundedLost-prospect<50[adj] Direct adjacency — agentic auth/integration. Their pitch overlaps; their customers may be Aileron prospects. NewLinkedIn ↗
ComposioYC/FundedLost-prospect50-100Direct competitor on the agent-tool-integration axis. Lost-prospect: target their customers, not Composio itself. NewLinkedIn ↗
Arcade AIYC/FundedLost-prospect<50[adj] Agent tool catalog; direct competitor adjacency. Same advice as Composio. NewLinkedIn ↗
Lindy AIYC/FundedBuyer<50[adj] AI assistant builder; their power users connect production services and feel the credential pain. NewLinkedIn ↗
Sema4.ai (was Robocorp)YC/FundedBuyer100-200Agent platform with enterprise audience; mature pain around audit and credentials. NewLinkedIn ↗
PineconeYC/FundedPractitioner200-400Vector DB; their customers run agents in production. Peer perspective on what those customers ask for. NewLinkedIn ↗
WeaviateYC/FundedPractitioner50-150Vector DB; same as Pinecone. NewLinkedIn ↗
StripeJob-signalBuyer500-2000[adj] Larger. Active 'AI platform' hiring; financial compliance forces audit/RBAC to be load-bearing. NewLinkedIn ↗
ShopifyJob-signalBuyer1000-2000[adj] Heavy AI investment; Sidekick agent; engineering platform sophistication. NewLinkedIn ↗
AirbnbJob-signalBuyer1000-2000[adj] AI platform team mature; multi-tenant safety pressure. NewLinkedIn ↗
NotionJob-signalBuyer200-500Notion AI runs on Anthropic Managed Agents (Apr 2026) — but that's ONE product surface. Internal eng/support/ops agents are not. Platform team is the cross-cutting buyer; AI maturity + budget already proven. Anchor outreach: 'I saw Notion AI launch on Managed Agents — what's the story for everything else?' source ↗NewLinkedIn ↗
LinearJob-signalBuyer50-150AI features rolling out; small disciplined team; design-partner-friendly. NewLinkedIn ↗
CodaJob-signalBuyer100-300AI features in product. Similar to Notion shape. NewLinkedIn ↗
AtlassianJob-signalBuyer1000-2000[adj] Atlassian Intelligence + Rovo on Managed Agents — but Atlassian's eng platform runs many more agent surfaces (Jira, Confluence, Bitbucket automations, internal SRE). Platform team is the cross-cutting buyer; the Managed Agents decision proves AI fluency. source ↗NewLinkedIn ↗
FigmaJob-signalBuyer500-1000[adj] AI in design tools + Make; strong platform culture. NewLinkedIn ↗
CanvaJob-signalBuyer1000-2000[adj] AI features at scale. NewLinkedIn ↗
LoomJob-signalBuyer100-300AI summarization in product; Atlassian-owned, smaller team unit. NewLinkedIn ↗
PlaidJob-signalBuyer500-1000[adj] Fintech compliance; AI assistant builds happening. NewLinkedIn ↗
BrexJob-signalBuyer300-600AI assistant in product; financial-services compliance posture. NewLinkedIn ↗
RampJob-signalBuyer300-600Same as Brex; aggressive AI feature build. NewLinkedIn ↗
MercuryJob-signalBuyer100-300Banking-adjacent; AI feature push; appropriate size. NewLinkedIn ↗
DatadogJob-signalBuyer1000-2000[adj] Bits AI agent in product; mature platform org. NewLinkedIn ↗
PagerDutyJob-signalBuyer500-1000[adj] AI Ops agent; incident-management context forces audit clarity. NewLinkedIn ↗
HashiCorpJob-signalBuyer1000-2000[adj] AI features integrated; secrets-mgmt heritage means they grok the credential story instantly. NewLinkedIn ↗
ConfluentJob-signalBuyer1000-2000[adj] AI platform investment; data-pipeline shape. NewLinkedIn ↗
SnowflakeJob-signalBuyer2000+[adj/oversized] Cortex agents. Sales cycle slow but signal exists. NewLinkedIn ↗
DatabricksJob-signalBuyer2000+[adj/oversized] Heavy agent investment. Same caveat as Snowflake. NewLinkedIn ↗
dbt LabsJob-signalBuyer200-400AI in dbt Cloud; data-platform audience. NewLinkedIn ↗
HightouchJob-signalBuyer100-200Reverse-ETL + AI; agents acting on consequential systems by definition. NewLinkedIn ↗
CensusJob-signalBuyer50-150Similar to Hightouch. NewLinkedIn ↗
TemporalJob-signalPractitioner100-200Workflow infra; their customers run agentic workflows. Peer perspective valuable. NewLinkedIn ↗
RenderJob-signalPractitioner50-150PaaS; sees agent runtime traffic shape from customers. NewLinkedIn ↗
Fly.ioJob-signalPractitioner50-100Same as Render. Strong agent-runtime customer base. NewLinkedIn ↗
RailwayJob-signalPractitioner<50[adj] PaaS; smaller; agent-friendly community. NewLinkedIn ↗
RedditJob-signalBuyer1000-2000[adj] Public AI investment push 2025-2026. NewLinkedIn ↗
DoorDashJob-signalBuyer1000-2000[adj] ML platform mature; agent expansion typical for ops-heavy biz. NewLinkedIn ↗
InstacartJob-signalBuyer1000-2000[adj] Same shape as DoorDash. NewLinkedIn ↗
HubSpotJob-signalBuyer2000+[adj/oversized] AI tooling buildout; enterprise sales cycle. NewLinkedIn ↗
VantaDesign-ptnrBuyer300-600Compliance automation co; SOC 2 audit-trail thinking is native. Strong instinct for the v4 story. NewLinkedIn ↗
DrataDesign-ptnrBuyer200-400Compliance peer of Vanta. NewLinkedIn ↗
SecureframeDesign-ptnrBuyer100-250Same compliance posture; AI agent expansion underway. NewLinkedIn ↗
HoneycombDesign-ptnrBuyer100-200Observability; shipping AI features. Design-partner-friendly culture. NewLinkedIn ↗
SentryDesign-ptnrBuyer300-600Sentry Seer (their product feature) runs on Managed Agents — but their internal SRE team runs agents for incident triage, runbook automation, on-call. That's exactly the consequential-systems workload v4 governs. Strong design-partner culture too. source ↗NewLinkedIn ↗
TinesDesign-ptnrLost-prospect100-200Security automation + workflow + AI; conceptual overlap. Their customers are the targets. NewLinkedIn ↗
WizDesign-ptnrBuyer500-1000[adj] Cloud security; AI features in product. Security buyer well-formed. NewLinkedIn ↗
SnykDesign-ptnrBuyer1000-2000[adj] Security + DeepCode AI agents. NewLinkedIn ↗
Orca SecurityDesign-ptnrBuyer300-600Cloud security; AI investment. NewLinkedIn ↗
Cal.comDesign-ptnrBuyer<50[adj] Smaller; OSS-native; community-friendly; loves design-partner relationships. NewLinkedIn ↗
ResendDesign-ptnrBuyer<50[adj] Smaller; agent-friendly infra co; tight founder network. NewLinkedIn ↗
Eight SleepDesign-ptnrBuyer100-200Vercel customer; AI in product; design-partner culture. NewLinkedIn ↗
SonosDesign-ptnrBuyer500-1000[adj] Larger consumer; AI features; Vercel/peer-infra customer. NewLinkedIn ↗
Hugging FaceDesign-ptnrLost-prospect100-300Infisical customer (public). Almost certainly built agent infra internally; lost-prospect with strong opinion data. NewLinkedIn ↗
Hex (data team)Design-ptnrPractitioner100-200Listed under YC too; their data team specifically is a practitioner target. NewLinkedIn ↗
RetoolDesign-ptnrBuyer300-600Retool Agent in product. Agents acting on consequential systems by definition for customers. NewLinkedIn ↗
Airplane / PylonDesign-ptnrBuyer50-150Workflow/runbook adjacent. Verify size. NewLinkedIn ↗
Vercel customers (case studies)Design-ptnrBuyervariedSource from vercel.com/customers; each is pre-warmed to infra-product design-partner asks. NewLinkedIn ↗
Cloudflare Workers AI customersDesign-ptnrBuyervariedSource from cloudflare.com case studies; agent-runtime adjacent. NewLinkedIn ↗
Modal customers (case studies)Design-ptnrBuyervariedSource from modal.com/use-cases — peer-infra customers comfortable with this motion. NewLinkedIn ↗
KlarnaLost-prospectLost-prospect1000-2000[adj/oversized] Publicly built/replaced workforce with AI agents. The poster child. They're not buying; they're the case study for talking to peers who saw it. NewLinkedIn ↗
Block (Goose framework)Lost-prospectLost-prospect2000+[adj/oversized] Built and open-sourced Goose. They chose to roll their own. Strongest possible 'why we didn't buy' data. NewLinkedIn ↗
Cognition AI (Devin)Lost-prospectLost-prospect100-300Competitor + lost-prospect. Their customers' decision criteria are gold for the Fleet (Option A) thesis specifically. NewLinkedIn ↗
Magic.devLost-prospectLost-prospect50-150Building own agent + substrate. Same as Cognition. NewLinkedIn ↗
AsanaJob-signalBuyer1000-2000[adj] Asana Intelligence on Managed Agents — but their platform org runs many internal agents (project automations, internal tools, ops). Multi-deployment buyer story; AI fluency proven by the Managed Agents adoption. source ↗NewLinkedIn ↗
Monday.comLost-prospectLost-prospect1000-2000[adj] Same as Asana. NewLinkedIn ↗
ClickUpLost-prospectLost-prospect300-600AI features; building own. NewLinkedIn ↗
IntercomLost-prospectLost-prospect500-1000[adj] Fin agent shipped; built their own. The reasons matter. NewLinkedIn ↗
DriftLost-prospectLost-prospect200-500Conversational AI for sales; built their own substrate. NewLinkedIn ↗
GongLost-prospectLost-prospect1000-2000[adj] AI for sales; enterprise compliance posture. Built their own. NewLinkedIn ↗
SalesloftLost-prospectLost-prospect500-1000[adj] Same shape as Gong. NewLinkedIn ↗
OutreachLost-prospectLost-prospect500-1000[adj] Same shape. NewLinkedIn ↗
LiltLost-prospectLost-prospect100-300Translation AI agents in production. Probably built their own; strong opinion data. NewLinkedIn ↗
Vellum AILost-prospectLost-prospect<50[adj] LLM tooling — direct competitor adjacency. Find their customers via case studies. NewLinkedIn ↗
HumanloopLost-prospectLost-prospect<50[adj] LLM observability; partial competitor. NewLinkedIn ↗
Arize AILost-prospectLost-prospect100-200ML/AI observability; their customers chose Arize over building. Adjacent decision criteria. NewLinkedIn ↗
Salesforce Agentforce customersLost-prospectLost-prospectvariedCustomers who chose Agentforce over rolling their own — interview a few via Salesforce case studies. Tells you the SaaS pull. NewLinkedIn ↗
OpenAI Agent Builder usersLost-prospectLost-prospectvariedPublic users of OpenAI Agents SDK in production. Tells you why they didn't need credential mediation outside the SDK. NewLinkedIn ↗
Anthropic Managed Agents customersLost-prospectLost-prospectvariedHardest signal — these chose the most direct alternative to v4. Reach via Anthropic case studies + community posts. NewLinkedIn ↗
Infisical Agent Vault (when GA)Lost-prospectLost-prospectvariedCurrently a research preview (per Infisical comms, May 2026) — not production-ready. Watch for the commercial GA and its first customer disclosures; until then there's no public lost-prospect pool to mine. source ↗NewLinkedIn ↗
Cloudflare (iMARS / Agents)Lost-prospectLost-prospect2000+[adj/oversized] Cloud 2.0 / agentic cloud announcement 2026. Internal iMARS tiger team built their own AI engineering stack on top of Workers. Lost as customer; richest peer-practitioner conversation. source ↗NewLinkedIn ↗
Microsoft (Copilot Studio team)Lost-prospectLost-prospect2000+[adj/oversized] Build their own. Worth one conversation for strongest 'why-not-Aileron' signal at scale. NewLinkedIn ↗
Slack engineering (Salesforce)Design-ptnrBuyer1000-2000[adj/oversized] Slack-specific eng platform; AI agent expansion. Long sales cycle but rich practitioner conversations. NewLinkedIn ↗
GitHub (Copilot Workspace)Lost-prospectLost-prospect2000+[adj/oversized] Built own substrate. Why-not data + practitioner perspective. NewLinkedIn ↗
PostmanJob-signalBuyer300-600Postbot AI agent; API-first culture aligns with Aileron's developer audience. NewLinkedIn ↗
MuxDesign-ptnrBuyer100-200Video infra; design-partner-friendly culture; AI features expanding. NewLinkedIn ↗
WorkOSDesign-ptnrBuyer50-150Auth/SSO infra; founder network overlap; deeply understand the identity-at-the-agentic-layer message. NewLinkedIn ↗
ClerkDesign-ptnrBuyer50-150Auth as a product; understands the identity message immediately. NewLinkedIn ↗
KnockDesign-ptnrPractitioner<50[adj] Notification infra; agent-friendly customer base; tight design-partner muscle. NewLinkedIn ↗
InngestYC/FundedPractitioner<50[adj] Workflow infra for agents; direct adjacency — their customers are runtime-aware. NewLinkedIn ↗
Trigger.devYC/FundedPractitioner<50[adj] Workflow infra; similar to Inngest. NewLinkedIn ↗
DeferYC/FundedPractitioner<50[adj] Background-job infra; agent-relevant. NewLinkedIn ↗
LiveblocksYC/FundedPractitioner<50[adj] Realtime infra; AI features expanding. Founder-network adjacent. NewLinkedIn ↗
StytchDesign-ptnrBuyer100-200Auth infra; grasps identity-at-agent-layer message immediately. NewLinkedIn ↗
KnapsackYC/FundedPractitioner<50[adj] AI ops infra; small but plugged-in. NewLinkedIn ↗
PulumiJob-signalBuyer100-300Pulumi AI in product; IaC-adjacent culture grasps runtime mediation quickly. NewLinkedIn ↗
Bytebot / Operator-style agent usersLost-prospectLost-prospectvariedCompanies running browser-driven agents in production. Find via Browserbase customer list, OpenAI Operator case studies. NewLinkedIn ↗
WonderfulLost-prospectLost-prospect200-400Amsterdam-based; $150M Series B March 2026 ($2B valuation). 30+ countries; production-grade agents in telecom/finserv/manufacturing/healthcare. McKinsey partnership. Direct competitor at enterprise scale. source ↗NewLinkedIn ↗
InsightFinderYC/FundedPractitioner50-100$15M Series B April 2026 (Yu Galaxy). AI agent observability/reliability — agent eval, deploy, monitor. Peer/adjacent on the observability axis. source ↗NewLinkedIn ↗
RakutenJob-signalBuyer2000+[adj/oversized] Multi-business conglomerate using Anthropic for one slice (Managed Agents + Agentic Coding). Many other agent surfaces across e-commerce/finance/telecom/content businesses; platform team is the cross-cutting buyer. source ↗NewLinkedIn ↗
PendoJob-signalBuyer500-1000[adj] Recent Anthropic Managed Agents case study (May 8, 2026). Pendo AI feature on Managed Agents; their internal eng + customer-facing automation are separate surfaces. Fresh AI-decision context to anchor outreach. source ↗NewLinkedIn ↗
TELUSJob-signalBuyer2000+[adj/oversized] In Anthropic 2026 Agentic Coding Trends Report. Telecom with strict compliance posture; their AI Platform team is responsible for governance across many agent surfaces beyond coding. source ↗NewLinkedIn ↗
ZapierJob-signalBuyer500-1000[adj] In Anthropic 2026 Agentic Coding Trends Report (internal eng usage). Zapier's *product* is workflow automation; their *internal platform* governs the agents that build/operate it. Cross-cutting buyer profile. source ↗NewLinkedIn ↗
CREDJob-signalBuyer500-1000[adj] In Anthropic 2026 Agentic Coding Trends Report. Indian fintech, strict compliance; useful counterpoint to US-only ICP framing. Platform team is the cross-cutting buyer for compliance/audit. source ↗NewLinkedIn ↗
BMOJob-signalBuyer2000+[adj/oversized] Anthropic financial-services template early deployer. Regulated buyer is exactly the B-positioning archetype; AI fluency + budget already proven. Sales cycle slow but interview signal high. source ↗NewLinkedIn ↗
Amalgamated BankJob-signalBuyer500-1000[adj] Anthropic financial-services template early deployer. Smaller bank than BMO; potentially closer to ICP shape. Same regulated-buyer signal. source ↗NewLinkedIn ↗
NubankJob-signalBuyer2000+[adj/oversized] Devin in production for monolith refactoring — but that's code-gen, one use case. Nubank's customer-facing fintech ops run many other agents; platform team is the cross-cutting buyer. Fintech compliance = B-positioning archetype. source ↗NewLinkedIn ↗
VismaJob-signalBuyer2000+[adj/oversized] Devin + Copilot + Azure combo for code. Their accounting/business-software product line has many AI surfaces beyond code-gen. Platform/security team buys for cross-cutting governance. source ↗NewLinkedIn ↗
Dell TechnologiesJob-signalBuyer2000+[adj/oversized] Devin for code-gen. Massive enterprise with many independent agent initiatives across product lines. Platform/security team is the buyer; sales cycle slow. source ↗NewLinkedIn ↗
Cisco SystemsJob-signalBuyer2000+[adj/oversized] Devin for code-gen. Same shape as Dell. source ↗NewLinkedIn ↗
CognizantDesign-ptnrPractitioner2000+[adj/oversized] Cognition strategic partnership — they deploy Devin into clients. Not a direct buyer; potential channel partner. Ask their FDE leads what runtime-layer pain shows up across their portfolio. source ↗NewLinkedIn ↗
InfosysDesign-ptnrPractitioner2000+[adj/oversized] Cognition strategic partnership; rolling Devin out in FS practice. Channel-partner shape, not direct buyer. Strong sourcing channel for regulated-FS contacts. source ↗NewLinkedIn ↗
11xJob-signalBuyer50-150Composio customer (tool integration) for their AI sales agents Alice + Mike. Composio gives them connectors; it doesn't give them runtime sandbox, policy, audit, or credential mediation. ICP-sized; complementary-not-competing pitch is direct. source ↗NewLinkedIn ↗
SwarmZeroJob-signalBuyer<50[adj] Composio customer; multi-agent orchestration. Tool integration layer ≠ runtime; complementary pitch. source ↗NewLinkedIn ↗
Assista AIJob-signalBuyer<50[adj] Composio customer; rapid integration. Same complementary framing. source ↗NewLinkedIn ↗
C.H. RobinsonJob-signalBuyer2000+[adj/oversized] LangSmith customer (observability) for logistics agents — 5,500 orders/day, 600+ hours/day saved. LangSmith gives them traces; doesn't give them runtime/policy/credential mediation. Complementary pitch; regulated logistics audit posture. source ↗NewLinkedIn ↗
ModaJob-signalBuyer<50[adj] LangSmith / Deep Agents customer; design-focused multi-agent. Complementary pitch. source ↗NewLinkedIn ↗
EY (agentic AI OS)Lost-prospectLost-prospect2000+[adj/oversized] Built their own with Microsoft + NVIDIA. Same Klarna-shaped 'we rolled our own' decision; interview for the why. source ↗NewLinkedIn ↗
Clawvisor customersLost-prospectLost-prospectvariedClosest direct competitor on credential mediation. 14 services curated; supports Claude Code, OpenClaw, Hermes, Perplexity Computer. Find their public users via case studies / Twitter. source ↗NewLinkedIn ↗
Anthropic Managed Agents users (beyond launch list)Lost-prospectLost-prospectvariedSource via the Customers page; new case studies posting monthly. Watch for ICP-shaped (50-500 eng) customers especially. source ↗NewLinkedIn ↗
OpenAI Frontier early adoptersLost-prospectLost-prospectvariedOpenAI Frontier (Feb 2026) + OpenAI Deployment Company. Frontier is the most direct alternative to v4 at enterprise scale; their customers are the most informative lost-prospect set. source ↗NewLinkedIn ↗
ServiceNow Agentic AI customersLost-prospectLost-prospectvariedServiceNow + Accenture FDE program (May 2026) for moving agentic AI from pilot to production. Their customers chose 'enterprise platform + consultancy' over 'install a runtime.' Why? source ↗NewLinkedIn ↗
How this list was assembled
  • Refreshed from current web search: YC W26 batch, recent Series A/B/C announcements, Anthropic Managed Agents launch customers, Cognition/Devin enterprise rollouts, Composio + LangSmith case studies.
  • Where a row has a source ↗ link, that's where the claim came from. Verify before outreach if it's load-bearing.
  • Re-tiered after vendor-neutrality challenge. Anthropic Managed Agents customers (Notion, Atlassian, Asana, Sentry, Pendo, Rakuten, TELUS, Zapier, CRED, BMO, Amalgamated), Cognition/Devin enterprise rollouts (Nubank, Visma, Dell, Cisco), and Composio/LangSmith customers (11x, SwarmZero, Assista, C.H. Robinson, Moda) are now Buyers, not Lost. Why: vendor neutrality is the v4 pitch, and each of these orgs adopted an incumbent for ONE use case while running many other agent surfaces.
  • Lost-prospect is reserved for: (1) direct vertical-agent competitors (Sierra, Decagon, Cresta, Wonderful, Cognition itself, Intercom Fin); (2) public self-builders of an org-wide substrate (Klarna, Block/Goose, Cloudflare iMARS, EY, Microsoft Copilot Studio, GitHub Copilot Workspace); (3) competitor customer pools as sourcing channels.
  • Cognizant and Infosys moved to Partner. They distribute Cognition's Devin into clients, not direct buyers.
  • Size bands are estimates. [adj] = adjacent tier kept for ICP-edge testing. [adj/oversized] = useful interview signal but slow sales cycle.
  • "Find buyer" runs a Google query scoped to LinkedIn for likely titles. Refine for context.
  • Infisical Agent Vault is a research preview, no public production customers yet, so it's not a sourcing channel right now.
Outreach anchor for AI-mature buyers

For Anthropic / Devin / Composio / LangSmith customers, their public adoption is the anchor. AI fluency and budget already signaled, which makes this a warmer cold than a greenfield prospect. Use the V4 template at /outreach/.