Data and connections
Microsoft Teams, SharePoint documents, FAQs and product-system data bring knowledge into employees’ everyday workspace.
Questions in Teams that find their answers in company data
An AI assistant in Microsoft Teams answers employees' questions about procedures, people, absences and products, with a link to the source document and no answers from outside the company knowledge base.
Results
- 4 sources
- of knowledge in one conversation: FAQ, procedures, products, people
How it works
An employee asks in Teams
in plain language
The gateway picks the source
four routes
Searches the documents
SharePoint library
Answers with a link to the source
no answers from outside the knowledge base
How do I order a laptop for a new team member?
Question about an IT procedure
FAQProceduresProducts (PIM)People and absences
IT procedure · ordering equipment
Scope of the procedure
Request in the IT portal, “Equipment” category
Manager approval
Pickup and handover
passage foundHow do I order a laptop for a new team member?
Submit a request in the IT portal under “Equipment”. Your manager needs to approve it.
IT procedure · ordering equipmentanswer with a source link
Result4 sourcesof knowledge in one conversation: FAQ, procedures, products, people
01
Challenge
Knowledge about IT, HR and admin procedures, the company structure and products sat in SharePoint, Excel, OneNote, the PIM system and in the heads of individual people. As a result, questions went to support teams and IT. Previous solutions were connected neither to Teams nor to product data.
Employees need an answer and the source that lets them verify it.
02
Solution
In 14 weeks we took the assistant to production in Teams. In a single conversation it uses a base of frequently asked questions, company procedures, product data from the PIM, and the absence list and company structure, and it gives answers from documents with a link to the source. Before release it was tested by IT, testers from various departments and the product support team. An analysis of the first messages showed that 40% of questions concern products, so these are to go to a central product database being built in a separate project.
03
How we built it
A message from Teams goes through Azure Bot Service to an n8n workflow, where a gateway classifies the question and picks a path: a ready answer from the FAQ, document search, product data or the absence list. Documents from a dedicated SharePoint library are converted to text and indexed in a vector database along with a link to the source file. Typical questions get a ready answer from the FAQ, with no generation and no document search, which lowers response cost and time. Every message and rating goes to a log with usage views, and testers evaluate new versions in a separate environment before they reach Teams.
Microsoft TeamsAzure Bot ServiceSharePointn8nQdrantSupabaseClient's PIM
Workflow
How the solution works
- 01
Ask in Teams
An employee asks in their everyday tool without first choosing a folder or system.
- 02
Get an answer with sources
The assistant uses FAQs, documents and product data, pointing to the materials behind its answer.
- 03
Review and improve
Questions and feedback reveal knowledge gaps. Unclear topics are passed for review and updates.
The human role
Information owners verify the content. The assistant helps find and explain it; it does not replace the responsible person’s decisions.
Lessons
What made the difference.
- 01
An answer with a link to the source
An answer based on documents points to the source file, and the link opens only on a company account.
- 02
A gateway before the model
A question first goes to classification: typical ones get a ready answer from the FAQ, and questions outside company topics are rejected.
- 03
Operational data, not just documents
The assistant checks the absence list and company structure through filtered queries, and pulls product data from the PIM system.
Knowledge agentRAGTeams integration
Company of the Future
How this builds the Company of the Future
A project delivers results on its own, but it also leaves organized context behind in the company. Every next AI solution builds on it. This project added to the AI Engine:
- Context Lakehouse
FAQ, procedures, product data from the PIM, absences and the company structure available in one conversation, with a link to the source.
- MCP
Teams, SharePoint and the PIM connected through n8n: the gateway decides where the answer comes from.
- Decisions
The question log shows what people look for: 40% of questions are about products, so they are to go to a central product database built in a separate project.
Value
Where value is created
Time saved
Less searching for documents and repeating common questions.
New possibilities
Shared knowledge supports more topics without relying on the same experts.
What comes next
New questions help prioritise missing information and the next sources to connect.
Most of our projects are covered by NDAs, so we only share part of them, without client names or details, as inspiration. We are happy to tell you more about our experience in a conversation.
Let’s talk about the work you want to improve.
Tell us how a process works today. Together, we will explore where AI Engine can help and which module to start with.
Book a callOr write to: hello@wearefuture.ai
