Skip to content
BobSentry

Product

Runtime authorization before consequential AI-agent actions

BobSentry is not an orchestration framework, observability platform, chatbot guardrail, workflow automation product, or generic AI governance dashboard. It evaluates whether a specific action should proceed right now—and returns ALLOW, REQUIRE_APPROVAL, or BLOCK.

Agent requests consequential action

Runtime

BobSentry evaluates policy

Decision

ALLOW

ALLOW

Policy permits the requested action

REQUIRE_APPROVAL

REQUIRE_APPROVAL

Human authorization required before proceed

BLOCK

BLOCK

Policy denies the requested action

Authorization

Human authorization where required

Execution

Integrator executes when authorized

Evidence

Signed decision evidence

Authorization before execution. BobSentry does not execute the downstream action—the integrator does, only when policy allows.

Runtime decision contract

Every consequential evaluation returns one of three decisions.

ALLOW

ALLOW

Policy permits the requested action.

REQUIRE_APPROVAL

REQUIRE_APPROVAL

Human authorization is required before the integration should proceed.

BLOCK

BLOCK

Policy denies the requested action.

Human authorization for the actions that matter

Humans should not approve every AI action. They should approve the actions that matter. When BobSentry returns REQUIRE_APPROVAL, an operator authorizes or rejects the specific proposed action—target, metadata, policy version, and expiration included. A standing permission is not enough. Approval authorizes the action; it does not execute it.

  1. 01Agent requests a consequential action
  2. 02BobSentry evaluates policy
  3. 03ALLOW / REQUIRE_APPROVAL / BLOCK
  4. 04Human authorization where required
  5. 05Integrator executes when authorized
  6. 06Signed evidence recorded

Authorization before execution. Only when policy permits—and human authorization is granted when required—should the integrator proceed.

Product surfaces

Operate Agents, Policies, Approvals, Evidence, and the Simulator in the tenant application.

Available

Agents

Register agent identities and issue Agent-bound runtime credentials that are evaluate-only and revocable.

Available

Policies

Define what may proceed automatically, what requires authorization, and what should be denied—including safe unmatched-action fallbacks.

Available

Approvals

Route consequential actions to a human when policy returns REQUIRE_APPROVAL. Approval authorizes the action; it does not execute it.

Available

Evidence

Keep signed authorization evidence: agent, action, policy, whether authorization was required, who authorized, and when. Execution outcomes are recorded separately from authorization.

Available

Simulator

Test how policy would decide for a proposed action without operational side effects.

Integration modes

Customer-facing modes today are COOPERATIVE and OBSERVE. There is no customer-selectable ENFORCED mode.

COOPERATIVE

The integration calls BobSentry before the consequential operation and is responsible for honoring the authorization decision.

OBSERVE

BobSentry evaluates and records what policy would decide, but does not control downstream execution.

Design-partner ready

If your agents can take consequential actions you are not ready to run fully autonomously, work with us as a design partner.

Become a design partnerView docsSecurity model