AI Agents That Know What They Are Allowed to Do
What Agents Do on the Platform
Agents research across multiple sources, summarize, review, create drafts and trigger follow-up steps – for example:
- memos and notes from case files
- responses to recurring inquiries
- compilations for audits and reviews
- preparing decisions
iHub provides micro-apps and assistants for recurring tasks and processes; specialized domain assistants are configured on the same foundation.
Why on the Platform?
On the Enterprise AI Platform, every agent works with the same permission-checked retrieval as search and chatbot – it sees only what the user is allowed to see and shows where its results come from.
That means for every agent:
- no separate permission concept
- no separate data integration
- no separate model setup
Governance for Agents
- Checked and logged: Every access step is checked against permissions and logged.
- Human-in-the-loop: Results can be submitted for review and approval before they take effect.
- Full control for administrators: They define which models, sources and tools an agent may use – including tools connected via MCP.
Your Own Agents
Agents you build with other frameworks or assistants can be supplied with the same permission-checked context via MCP and APIs.
This makes the platform the knowledge layer for agents not built by IntraFind as well.
💡FAQ - AI Agents
[APPROVAL Product Management: specify the scope of action capabilities (tool calling, workflow integration); only list what is available as standard]
Through the same permission checks as in search, configurable approvals per agent, and human-in-the-loop for results that take effect.
With the same gold standards and metrics as answers: groundedness, citation quality and retrieval accuracy – complemented by the traceability of each step in the audit trail.
Secure, Sovereign, Compliant
✔️EU AI Act - The Enterprise AI Platform complies with the AI Act (Regulation (EU) 2024/1689)
✔️Data protection - GDPR-compliant operation
✔️No model training - Your data is not used to train AI models
✔️Data residency - You decide where your data is processed and stored
We Look Forward to Hearing from You
Bring us a specific use case – a search that takes too long, a question you want a chatbot to answer, or a process an agent could take over.
We’ll show you how to build it on the platform – and what becomes possible next without starting over.
