Building the default independent safety layer for AI agents
Here's how we think about the market, and how we're going after it.
The investment thesis
Structural demand
Agents are moving from generating content to executing action. Enterprise safety controls have to keep pace.
A clear wedge
Enterprise buyers lack their own AI red team, and a vendor grading its own agent is a conflict of interest. Independent verification has a natural buyer.
A path to recurring revenue
Verify is the entry point to Guard and Sentinel Live, turning one-time project revenue into ongoing subscription.
A compounding moat
Every engagement strengthens the next. See the flywheel on our product page.
Land and expand
We land with a high-touch Verify engagement on a high-risk, permission-clear agent, then expand into Guard and Continuous Assurance using the threat model and trust built along the way.
Land
Enterprises actively deploying high-authority agents, starting in finance, procurement and internal tooling.
Expand
Verify findings become the basis for Guard and Sentinel Live, converting project revenue into subscription.
Team-led sales
Founders drive early customer discovery and technical delivery, building toward a repeatable sales motion.
From project revenue to safety infrastructure
| Stage | What the customer buys | Pricing logic |
|---|---|---|
| 1. First assurance | Sentinel Verify | Per project / per system |
| 2. Go-live | Sentinel Guard | Enterprise subscription + usage |
| 3. System change | Sentinel Live | Re-testing & annual service |
| 4. Ecosystem | Rating / standards / audit interfaces | Enterprise contracts, partnerships |
This business model, along with our product roadmap and milestones, reflects current startup hypotheses under active validation with customers, engineering practice, and standards bodies. It should not be read as fully realized product capability.