01
Agent visibility
Bring agent identities, owners, tools and observed behaviour into a shared inventory. Discovery coverage depends on the connected surfaces.
← FOR BUSINESS
AI agent security & governance
Your agents. Your rules. Inside your network.
On-premise security and governance for your AI agents. Keep the control plane, policy decisions and evidence inside your network, with an offline configuration using local models and internal systems.
CONTROL IN THE EXECUTION PATH
A policy boundary around agent actions.
Discover the agent, evaluate the request, enforce the verdict and retain the evidence. Supported calls can be allowed, sanitised or blocked before they reach the next system.
Visibility, enforcement and evidence.
02
Runtime policy
Evaluate governed prompts, responses and actions before dispatch. Allow, mask or block according to configured policies and supported integration coverage.
03
Sensitive-data protection
Run local sensitive-data checks before provider calls. Judge-based checks receive sanitised content; sanitised previews and audit metadata may still reach the configured control plane.
04
Tool and MCP controls
Apply rules to supported tools, shell operations, database queries and MCP invocations. Coverage and add-ons must be verified for each integration.
05
Behaviour and spend
Inspect changes in agent behaviour, policy alignment and model use, with cost visibility and configured limits.
06
Governance evidence
Review verdicts, matched policies and activity records. Evidence mapping supports reviews; it does not by itself establish certification or regulatory compliance.
DEPLOYMENT THAT FITS YOUR STACK
Connect where the work happens.
EgisAI documents in-process SDK governance, an inline gateway, self-hosted egress nodes and managed-browser discovery. Integration coverage, enforcement modes and package availability differ; we establish the right combination for your environment.
Agree the boundary before rollout.
- Identify agents, owners, tools and sensitive data.
- Check enforcement coverage and failure behaviour.
- Review control-plane location, retention and data flows.
- Validate policies before expanding enforcement.
Bring your agent environment into view.
Tell us which agents you run, what they can access and where you need stronger control. We will work through the integration and governance requirements with your team.
Talk to us about ACI SovereignTechnology sources and scope
Adapted from EgisAI product and self-hosting materials reviewed 23 October 2026. Offline semantic checks require a local model endpoint. External SaaS connectors and public model APIs require connectivity. Features vary by integration and package; no supplier certification is claimed for appana.
ON-PREMISE / ENTERPRISE & SOVEREIGN
Your infrastructure. Your intelligence. Your control.
ACI Enterprise and ACI Sovereign are scoped for on-premise deployment, with locally hosted models and an offline operating configuration.
Private by architecture
Keep knowledge, local inference and governance inside your network. Agree access, retention and operational controls with your team.
No external token bill
Locally hosted inference does not incur external model-token or API charges. Hardware, electricity, implementation and support still apply.
Room for more agents
Create agents around the workflows you need. Capacity is sized for your workloads, model choices and concurrent use.
Tuned to your work
Fine-tune specialists on approved examples and evaluate accuracy on unseen business tasks before expanding responsibility.
Offline operation requires local models and internal dependencies. Public model APIs, internet research and external SaaS connectors need connectivity when enabled.
CUSTOMERS & PRIORITY USE CASES
A runtime boundary between an agent intention and a real-world action.
Regulated and sensitive-data organisationsEnterprise AI and platform teamsSecurity, risk and internal audit teams
01 / Finance, healthcare administration and legal operations
Sensitive-data protection
Detect and mask configured personal data or secrets locally before governed model-based evaluation. Review the post-sanitisation evidence.
What to measure: Detection and masking quality, false positives and data-flow checks.
Explore the workflow02 / Enterprise AI platform teams
Agent tool and access boundaries
Restrict supported tool, MCP, database and shell actions to the operations each workflow is authorised to perform.
What to measure: Blocked forbidden actions, permitted-action success and bypass tests.
Explore the workflow03 / Teams connecting agents to untrusted content
Prompt-injection resistance
Test governed inputs and actions against deterministic rules and, where configured, a local semantic evaluator.
What to measure: Attack-test outcomes, false blocks and escalation handling.
Explore the workflow04 / Risk, compliance and internal audit
Evidence for control reviews
Trace an agent request to its policy verdict and the rule that fired. Export supporting records for the organisation’s own review.
What to measure: Trace completeness, reviewer effort and policy-change history.
Explore the workflow05 / AI operations and service owners
Behaviour and runaway-use monitoring
Observe model use, repeated calls and changes in behaviour. Review anomalous patterns and configured limits before extending agent access.
What to measure: Alert quality, unusual-loop detection and response time.
Explore the workflow06 / Finance and engineering operations
Governed financial or destructive actions
Apply configured rules before supported financial actions, destructive queries or shell commands are dispatched. Retain the verdict for investigation.
What to measure: Disallowed-action blocks, approval coverage and exception resolution.
Explore the workflowSovereign governs the agents carrying out these workflows. Enforcement, discovery and evidence coverage depend on the selected integration and configured policies.
Explore more use casesSEE THE WORKFLOW
Inspect the request. Enforce the rule. Keep the evidence.
01
Identify
Associate the governed call with its agent and execution context.
02
Inspect locally
Run deterministic rules and sensitive-data handling inside the customer environment.
03
Enforce before dispatch
Allow, sanitise or block according to the configured rule and supported call path.
04
Review the evidence
Record the verdict, matched policy and relevant context for the operator.