# Aevah Agent-Ready Platform

Aevah is an MCP-compatible tool host for enterprise agents. The same capabilities. The same governance. No privileged path.

Canonical page: https://www.aevah.com/platform/agents

## One surface, one set of rules

Every capability in Aevah — search, match, quality, stewardship, publication — is published as a versioned, schema-bound, permission-scoped tool that agents discover and call. Whether a person clicks it in the interface or an agent invokes it over MCP, it is the same capability under the same governance. There is no privileged path for agents.

## Structured for agents browsing on your behalf.

The site itself is agent-navigable. Aevah's documentation and capability information are exposed as WebMCP tools, so an agent browsing on your behalf can search the documentation, retrieve capability and security detail, and prepare a demo request for you to review — without scraping pages or guessing at structure. The Aevah application exposes the same kind of surface: an agent working alongside a person can search, inspect and propose, while anything consequential stays a human decision.

## Human decisions return a usable review surface.

When an agent reaches a decision only a person should make, Aevah hands back a usable interface rather than a wall of JSON. An agent proposing a merge returns a review surface showing the candidate records side by side, which attribute value wins and why, and approve or reject controls — rendered directly in the agent's own environment. The person decides on real evidence, in context, and the decision is captured in the audit trail. The interface is a rendering surface, never an authority: every approval is reauthorized against the approver's permissions before anything changes, and the surface never receives values the viewer is not cleared to see. Review surfaces cover merge and match review, survivorship explanation, data quality triage, mapping suggestion review, and generic proposal approval.

- Propose
- Review
- Approve
- Audit

## Three classes of tool.

An agent cannot promote itself between classes.

- Read: Returns data and changes nothing. Executes directly under the caller's permissions.
- Propose: Produces a change proposal for a human to approve. Nothing changes until a person with the required authority approves it. Most agent-initiated change belongs here.
- Execute: Performs a governed change under explicit prior confirmation, reauthorized at the moment it runs. Confirmation is single-use, bound to the specific operation, and cannot be reused or broadened.

## The enterprise capability surface.

Aevah exposes these capability families through the governed agent surface.

- Catalog discovery: Returns the tools, versions, classifications, and permissions available to the caller.
- Data navigation: Returns authorized entities, relationships, definitions, and business context.
- Quality intelligence: Returns quality standards, failed checks, affected records, and supporting evidence.
- Matching and survivorship insight: Returns candidate records, comparisons, proposed outcomes, and the evidence available for review.
- Governance memory: Returns definitions, policies, ownership, decisions, approvals, and audit history.
- Publication and activation: Returns publication status, eligible targets, activation outcomes, and correlation identifiers.
- Operations: Returns operation status, progress, cancellation state, and per-record outcomes.

## Agents inherit your rules, not new ones.

Every tool call carries the identity of whoever triggered it and can never exceed that identity's permissions. Sensitivity classification is honored on the way out, so a tool returns only what its caller is cleared to see. Actions that change data are recorded as new versions rather than overwrites, are safe to retry without duplicating work, and leave a complete audit trail — who acted, which tool, which version, what happened.

## Contracts don't move underneath you.

Tools are versioned. A breaking change publishes a new version; the existing one is never altered. Agents built against a supported version keep working exactly as before, and deprecation is signalled in advance with a replacement pointer.

## A governed change, end to end

This arc is the same whether the change originates with a person or an agent.

- Agent proposes
- Steward reviews in Aevah
- Approved and executed under confirmation
- New version and audit record

## Nothing is ever overwritten.

Governed changes are immutable and cumulative. Every change creates a new version; prior versions stay retrievable and comparable. Reverting moves forward — a prior version is reinstated by superseding the current one, so the record of what changed and what was undone survives intact. There is no destructive rollback, by design.

## Security and trust

Identity, authorization, classification, isolation, audit, and error behavior are documented at https://www.aevah.com/trust.

## Contact

- First Flight: https://www.aevah.com/how-aevah-works/start-small-and-expand
- Executive inquiries: mailto:executive@aevah.com
- Security inquiries: mailto:security@aevah.com