Cases
Case · International organisation · 245 employees · since 2024

One assistant. Different answers for different permissions.

In progress. Ongoing · since 2024

The challenge

Knowledge was spread across departments, procedures and systems. Getting the right information to the right person often depended on knowing who to ask, creating delays and bottlenecks.

What made it hard

Giving everyone access to the company’s knowledge while ensuring each person only sees what they’re allowed to see.

“What is the procedure for X?”
Your private AI
  • Employee General information
  • Department manager General information + Their department’s internal procedures
  • Executive General information + Authorised corporate information
Same question. Same assistant. Three permission contexts.

What we built

A central knowledge system, open to internal staff only and connected to the company’s knowledge, procedures and systems. Each person gets answers based only on the information they’re authorised to access.

Built on a controlled RAG architecture, integrated with internal systems through APIs and MCP, with models and data kept within the company’s controlled environment.

The principle

AI shouldn’t bypass the organisation’s access rules. It should understand them.

The assistant uses the same principle as the rest of the systems: people only access information they’re authorised to see.

Status

Currently being deployed. The first phase is the central knowledge system for internal staff. A personal assistant for each member will follow, along with the automation of processes such as support, information retrieval and calculations. Results will be published once adoption and operational impact have been measured.

Services and technology

More cases