Get in touch
← Back to ChatFox

Knowing what your AI can access should be straightforward.

Security and privacy are considered alongside the software itself. We define the information, permissions and operational controls for each project, then test them against the way the system will be used.

Discuss your requirements

Information used for a defined purpose

We identify which sources the solution needs, how they are connected and what information should remain out of scope. Your business data remains yours.

Technical detail

Approved source lists, retrieval configuration and scoped system credentials define the information available to each task.

Access appropriate to each role

Different people and AI clients may need different tools and information. We configure permissions around their responsibilities and the work they are authorised to do.

Technical detail

Authentication, role-based access and tool permissions are implemented for the project. Read and write access are assessed separately.

Controls around actions

We agree which steps can run automatically, which need validation and which require a person to approve. Incomplete requests and failures need a defined route for review.

Technical detail

The AI Control Layer exposes approved capabilities. Workflow rules and validation are applied around tool use; untrusted content must not grant new authority.

Clear handling and hosting arrangements

Hosting, processing, retention and export requirements are discussed and documented. Connected services are considered as part of that design.

Technical detail

ChatFox’s published hosting approach is UK-based for application data it hosts. Processing by connected services is assessed separately; project-specific arrangements define where information is handled.

Testing and accountability

We test realistic tasks and relevant failure cases with your team. Monitoring and audit requirements are agreed so important activity can be reviewed.

Technical detail

Checks can cover tool access, data validation, unexpected inputs, failed integrations and approval paths. Audit records and monitoring are configured for the agreed scope.

Ongoing management

We agree support responsibilities, maintenance and how changes will be reviewed as your business develops.

Technical detail

Connections, permissions, knowledge sources and application behaviour may need updating over time. Change review and support arrangements form part of delivery.

We draw on information-security and AI-governance control principles when designing our solutions. This is not a claim of ISO 27001 or ISO 42001 certification.

Tell us what you need your business to do better.

You don’t need a technical specification. Tell us about the process, problem or opportunity, and we’ll help define the right solution.