Neo’s sensor watches agentic software on your endpoints. aizome governs what every agent is doing, who it answers to, and what it may reach.
aizome is the Identity-first AI Governance Platform. It sits as a single secure exchange (AI Gateway) between every AI agent your workforce deploys and every enterprise system those agents access, governing identity, enforcing intent, controlling cost, and making every agent accountable from day one.
Neo is an agentic software control platform. Its endpoint sensor inventories AI agents, MCP servers, extensions, and apps, turning agentic attributes each action into a responsible user or agent, and enforces policy on tool calls before they complete.
Neo bet on the endpoint, because a sensor on the device is the vantage point from which you can intercept what an agent does. That bet is right about one thing: agents on employee machines were genuinely ungoverned, and Neo’s inventory of them is strong. aizome bet on the connection. An agent causes harm in the system it reaches, not on the laptop it runs from. Governing the connection means coverage follows the agent wherever it runs, and identity can be bound to the action rather than reconstructed from it afterward.
 |  | |
|---|---|---|
| Agent discovery | AI agents, MCP servers, extensions, and agentic apps across managed endpoints | Every agent connecting to an enterprise system: local, SaaS, containerized, MCP |
| Identity | Attributes each action to a responsible user, agent, application, or identity | Agents hold no credentials. Every action bound to a named employee through your IdP |
| Entitlement | Policies on tool calls, API access, and data movement | Evaluates the access behind the tool before the call executes |
| Intent | Compares runtime behavior against the software’s intended function | Every action checked against declared purpose before execution |
| Audit | Action-level audit trail tied to the responsible actor | Identity to tool call to outcome, tamper evident |
| Cost | Outside the endpoint model | Per-session token limits, cost drift detection |
| Human in the loop | Not publicly documented | Approval required before destructive actions |
| EU AI Act | Not structured for agent obligations | Articles 9, 12, 13, 14, 26 |
Agents hold no credentials of their own. Each action is bound to a named employee through the identity provider you already run.
Every action is compared with the purpose the agent declared, before it executes.
aizome evaluates what a tool can reach in the system it connects to, not only whether the agent may call it.
Neo is strongest where agents run on machines you manage: developer tools, local MCP servers, browser extensions, and the long tail of approved software that quietly turned agentic. Three questions decide whether that is sufficient.
Where do your agents run? A sensor reaches managed endpoints. Agents inside a SaaS platform, in a container, or stood up by a business unit in a workflow tool are outside that model.
Is attribution enough, or do you need identity? Attribution tells you afterward who was responsible. Identity decides beforehand what the agent may do, because it inherits a named person’s entitlements rather than holding its own credentials.
Who governs what the tool can reach? Controlling a tool call is not the same as knowing the access that tool carries on the other side, which is what sets the actual blast radius.
Thirty minutes, your environment. We will show you the agents nobody registered.
