Sign in with your work account
Employees use their work identity. Permissions are attached to the document and apply in the portal, AI tool and digital colleague.
Ownership & trust
Making company knowledge useful also requires clear agreements. About who can access it, where the data lives and what goes to the AI tool.
The data flow
These are three distinct places, each with its own role. Your choice of AI tool remains part of the assessment.
Documents you supply or connect through an agreed source. We define the scope and access during setup.
Runs in your own organisation’s Microsoft environment (Azure), at an address in your own domain. Documents are read and made searchable here. Fulkeno itself stores no customer data.
When answering a question, the tool can request permitted passages. That content goes to the chosen AI service and is also covered by your agreements with that supplier.
Fulkeno running in your own Microsoft environment does not automatically mean all data processing stays there. It also depends on the AI tool and contract you use.
Access follows the document
Employees use their work identity. Permissions are attached to the document and apply in the portal, AI tool and digital colleague.
Source material, AI interpretation and human records remain identifiable. An AI draft never silently becomes a human position.
Its own identity and limited permissions. For Deals, that means a dedicated mailbox and Teams channel, usage limits and an emergency stop.
Fulkeno does not use customer data to train models. Your team’s AI tools are additionally covered by your own supplier agreements.
What you build remains yours
Documents and recorded knowledge can be exported as ordinary files. Choosing another AI tool does not have to mean starting your company context again.
Yes, that is the starting point. Fulkeno runs in your own organisation’s Microsoft environment, and we set it up together with you. Your documents therefore do not sit with Fulkeno, and the address is in your own domain.
The connection has been used in Claude, Codex and Mistral. MCP is the open standard behind it. We assess and test support for another tool first; Copilot has not yet been investigated.
We assess that against the confidentiality terms, data flow and chosen AI service. The fact that Fulkeno runs in your own Microsoft environment does not, by itself, answer that question.
Deletion on request has a runbook. Execution with evidence needs to be revalidated on the current environment. Rehearsed recovery of agent memory is also a development item. We agree on required retention, deletion and continuity before use.
No. People deliberately record knowledge and reasoning. Workflow feedback is reviewed and can inform a measured release. Automatically deriving new workflows from use is a development direction.
Bring your IT partner. We review sources, access, AI tools and required agreements together.