AI Assistant & Agents.
Put AI to work on your data, within the rights you grant
Give customers a chat that answers only from content you have approved, and says so when it cannot answer rather than guessing. AI agents can work with your product too, under the same permissions and audit trail as any user. Usage limits keep costs predictable, and you choose the AI provider.
For your engineers.
What is included, in the terms your technical team will ask about.
Grounded chat that answers only from approved content, or declines with a reason
Agent access over the Model Context Protocol (MCP), using the same authorization as the API
Every agent action written to the audit log
Content from the public marked untrusted, to blunt prompt injection
Per-user and per-tenant request and token limits
Choice of provider: Anthropic, AWS Bedrock, OpenAI-compatible endpoints, or Gemini
Works with everything.
AI Assistant & Agents integrates seamlessly with other Functory capabilities and your existing tools.
Coda
Ready to ship?
We build with AI Assistant & Agents, so you do not have to.
Tell us what you are building, or browse the rest of the catalogue.