🌏 中文版
Funding Details
| Field | Value |
|---|---|
| Company | Comp AI (Miami, Florida, USA) |
| Round | Series A |
| Amount | $34M |
| Lead | Roo Capital, Grand Ventures (co-led) |
| Participants | No other participants disclosed |
| Valuation | Not disclosed |
| Total raised | $37.5M |
| Founded | 2025 |
| Headcount | Not disclosed (Grand Ventures says the team was 4 people at pre-seed) |
What the company does
Comp AI automates compliance and security work — having AI agents directly perform tasks that used to require manual effort, rather than just digitizing the paperwork around them.
Founded in 2025, the company's core pitch is that "first-generation compliance software just moved paper processes online; agentic AI can go much further." Its platform uses AI agents combined with organizational context to automate onboarding, policy and risk generation, evidence gathering, control monitoring, and vendor assessments — turning compliance from an annual checkbox exercise into continuous background work. This round also signals the company's next move: expanding from compliance automation into continuous cybersecurity, including real-time monitoring, control validation, and security testing across applications and infrastructure.
In under two years, annual recurring revenue has grown 15x year-over-year and customer count 10x, now exceeding 1,000 customers across healthcare, financial services, higher education, professional services, and software. Named customers include Corgi, Dub, OpenCode, Inference, and Primer.
What this round signals
What it means for the agent ecosystem
Comp AI's positioning reflects a broader shift: the next generation of compliance and security software competition isn't about whether audit workflows got digitized — it's about whether AI agents can actually understand business context and perform and verify the security work themselves. Co-founder Claudio Fuentes put it bluntly: compliance used to be only an approximation of security, because so much of the work could only happen manually and periodically. Agentic AI removes that constraint, letting work that once required a whole team happen continuously in software. This also mirrors a common pattern in enterprise agent adoption — starting with domains that have clear audit standards and highly repetitive workflows is an easier path to monetization than building a general-purpose agent from scratch.
What investors are betting on
Grand Ventures backing Comp AI from pre-seed — when the team was four people working out of a single room in New York — through to this round tells its own story: general partner Nathan Owen notes that a growing number of Grand Ventures' own portfolio companies have switched to Comp AI from rival platforms, effectively using real purchasing behavior inside its portfolio to validate the bet. Roo Capital's logic is more about timing: principal Varun Sridhar argues that AI is simultaneously changing software development speed and the threat landscape, making "continuous compliance" a requirement companies can no longer defer to an annual audit cycle.
Numbers worth watching
- 15x YoY ARR growth versus 10x YoY customer growth suggests most of the gain is coming from expansion within existing accounts, not just new customer acquisition
- Crossing 1,000 customers in under two years is a notably fast pace of customer acquisition for the compliance category
- Total funding of just $37.5M against 1,000+ customers gives Comp AI a markedly better capital-to-customer ratio than most agent startups still in the exploration phase
Watchlist status
Comp AI is not yet on the watchlist. Recommend adding it to section B7 (Agent Security / Governance) alongside Zenity, Noma Security, and Prompt Security, with tracking focus on: whether its agentic compliance platform can successfully expand from automating paperwork into real-time security execution, and how fast customers migrate away from incumbent compliance platforms.
Today's takeaway
Comp AI's round is a good reminder of something easy to overlook: the fastest commercial traction for agents often isn't in flashy general-purpose assistants — it's in work where the audit standard is already written down and the process is highly repetitive. That kind of task looks like it should have the lowest tolerance for agent error, but the clarity of the rules is exactly what makes it easiest to hand to an agent and still produce something verifiable, compared to open-ended tasks that require judgment calls.
References
Loading...