AgentRegistry
Correlates API-key usage, user and device approval, runtime registration, and AgentHook evidence to detect sanctioned AI use that is approved in principle but unmanaged in practice.
hookbusenterprise.com
HookBus Enterprise is the central governance platform for organisations running AI agents across teams, tools, models, and regulated workflows.
HookBus Agent starts governed locally. HookBus Enterprise centralises policy, approvals, runtime evidence, audit exports, subscriber health, and regulatory support.
Human-in-the-loop control is built into the execution path: Approve, Deny, or Ask before consequential actions continue.
Book a HookBus Enterprise demoSee HookBus AgentCompliance subscriber bundle
HookBus Enterprise packages the compliance subscriber bundle, Enterprise dashboard, commercial-use licence, support contract, and regulatory update packs that regulated deployments need.
Correlates API-key usage, user and device approval, runtime registration, and AgentHook evidence to detect sanctioned AI use that is approved in principle but unmanaged in practice.
Enterprise policy enforcement and execution control. Blocks, asks, approvals, denials, override paths, and runtime restrictions attached to real tool activity.
Runtime evidence, audit records, exportable trails, and retention support for assurance programmes.
Human-in-the-loop approval routing for Approve, Deny, and Ask decisions, with durable approval state and two-person verification where required.
Alerts, escalation paths, incident-shaped event detection, and regulator-format reporting workflows.
Controlled context and knowledge ownership. Keeps KB and context governance out of the enforcement subscriber.
Approved AI access is not the same as governed AI activity.
Enterprise platform
The open HookBus transport remains the runtime event layer. HookBus Enterprise adds the commercial governance plane enterprises need when agents move from experiments to production-critical workflows.
Architecture
Compliance posture
HookBus Enterprise is built for deployments where a CISO signs a DPIA before the tool goes live. Regulated AI programmes increasingly need runtime evidence: prompts, model calls, tool calls, policy decisions, human approvals, denials, and audit trails.
Supports automatic event-logging and audit-trail programmes by producing runtime evidence tied to agent actions.
Supports human oversight patterns with ask, approve, deny, edit-before-approve, interrupt, and override paths.
Produces event-level evidence for internal controls, assurance work, and audit preparation.
Supports incident and operational resilience workflows through evidence capture and notification patterns.
Security posture documentation, deployment notes, and customer assurance support available under NDA.
HookBus Enterprise supports customer assurance programmes. It does not certify compliance on its own.
Regulatory coverage mapping
The product is not sold as a magic compliance certificate. It is sold as runtime evidence and control infrastructure. These mappings show the control surface HookBus Enterprise gives a customer compliance programme.
Who buys this
Every buyer role has a different problem. HookBus Enterprise gives them one runtime evidence layer instead of five disconnected control projects.
Know every consequential action every agent takes, and stop actions before they execute when policy requires it.
Keep regulated data inside the approved boundary and prove what context was used in each decision.
Write policy once and apply it across models, agents, coding assistants, workflow runners, and internal tools.
Export decision trails, approvals, overrides, denied actions, model calls, tool calls, and evidence packs on demand.
Translate regulatory obligations into runtime controls that generate evidence automatically.
Keep teams moving with governed execution instead of waiting for every AI use case to become a bespoke approval project.
LLM strategy
HookBus Enterprise sits at the hook and evidence layer, not the model layer. Your models keep running wherever the enterprise has standardised.
AWS Bedrock, Azure OpenAI, Google Vertex AI, IBM watsonx.ai, or approved enterprise inference endpoints.
Ollama, vLLM, llama.cpp, NVIDIA NIM, Granite, Qwen, Mistral, or local model gateways.
OpenAI, Anthropic, Google, Mistral, DeepSeek, Qwen, or other providers routed through customer-approved controls.
Deployment posture
The old Agentic Thinking page had the right deployment language. This is the cleaned-up version for the dedicated product site.
Runs inside the customer's network with customer-controlled storage, policy, identity, and audit boundaries.
Deploy into AWS, Azure, or GCP private networks without making the governance plane a public SaaS dependency.
Support for environments where no outbound dependency is acceptable for policy decisioning or evidence storage.
Events, policy context, approvals, audit chain, and reporting data stay where the customer puts them.
Runtime and system integrations
The integration story from Agentic Thinking belongs here because it answers the enterprise buyer's first technical question: what does this actually connect to?
Claude Code, Cursor, Amp, GitHub Copilot, Hermes, OpenClaw, Codex, Anthropic Agent SDK, OpenAI Agents SDK, and HTTP-capable runtimes.
AgentProtect can combine deterministic policy, customer policy packs, and customer-approved semantic review.
ServiceNow, Jira, GitHub, GitLab, Azure DevOps, Slack, Teams, email, and escalation channels.
SIEM, DLP, audit systems, compliance repositories, IAM context, and enterprise evidence stores.
Intellectual property
HookBus Enterprise is built on Agentic Thinking's patent-pending runtime governance architecture for AI agents and action-level enforcement.
Patent application reference: GB2608069.7. Covers the HookBus runtime evidence and governance architecture.
Patent application reference: GB2604445.3. Covers the Composable Rule Enforcer and action-level policy enforcement architecture.
HookBus is a trademark of Agentic Thinking Limited. Third-party names are used nominatively to describe integration and do not imply affiliation or endorsement.
Built by Agentic Thinking Ltd
HookBus Enterprise is a commercial product from Agentic Thinking Ltd, a UK company building runtime governance infrastructure for AI agents in regulated organisations.
Leo Ruocco has 27 years across UK financial services, insurance, and defence, delivering complex programmes where controls, evidence, auditability, and operational discipline matter.
The product is not a thin AI wrapper. It is a governance architecture for organisations that need policy decisions, human-in-the-loop Approve / Deny / Ask controls, runtime evidence, and audit exports before agents touch real systems.
Company, trust posture, product family, and enterprise background.
Open transport layer and developer documentation.
Commercial governed agent app with local HookBus Core built in.
Open runtime event and evidence standard.
Agentic Thinking Limited does not claim that buying software alone makes an organisation compliant. HookBus Enterprise generates runtime evidence and controls that support customer compliance, assurance, and audit programmes.
Partnerships
AI runtime, coding assistant, and agent framework companies can license HookBus Enterprise for white-label embedding, OEM resale, or joint engineering. Your product gets an enterprise governance story your customers can buy.
Ship HookBus Enterprise under your brand with the bus, subscribers, policy packs, and evidence model.
Resell HookBus Enterprise inside your own paid enterprise tier.
Co-develop publisher shims so your runtime emits HookBus-compatible events.
Next step
The demo should show one real governed workflow: an agent tries a consequential action, HookBus Enterprise evaluates the action, a policy decision is made, approval is routed if needed, and the evidence trail is exported.
Book a 30-minute demoEmail sales