Knowledge Assistants
AI Company Memory
Captures decisions, context and lessons from meetings and threads into searchable memory.
Knowledge Assistants
AI Company Knowledge Hub for Microsoft 365
SharePoint, OneDrive, Outlook and Teams content answered with a citation, inside the Microsoft Graph permissions your admin grants.
Most Microsoft 365 companies have the answers already, spread across SharePoint sites, OneDrive folders, Outlook mail and Teams channels. This hub connects through Microsoft Graph and answers questions with a link to the source, right inside Microsoft Teams. Your IT team decides exactly which permissions it gets, and your files never leave your tenant's control.
- Ask from Microsoft Teams and get the source linked
- Finds the current SharePoint or OneDrive version for you
- IT approves every Microsoft Graph permission before launch
$12,400
20–30 working days
Optional maintenance $620/mo
What it is
AI Company Knowledge Hub for Microsoft 365 is a cited Q&A assistant connected natively to Microsoft 365 through Microsoft Graph, reading and updating SharePoint, OneDrive, Outlook and Teams data within approved scopes. Staff can ask from Microsoft Teams, and each answer links to its source. IT controls every Graph permission.
Why businesses buy it
SharePoint sprawl is expensive. People cannot find the current policy, so they ask a colleague, use an old copy or recreate a document that already exists. This hub answers from your Microsoft 365 content with the source attached, in the Teams window staff already have open, so the right version gets used and fewer questions interrupt your experts. The Scope Card fixes the Graph permissions, sources, document volume and testing in writing before you order, so IT and security can sign off on a known, limited connection.
Problems this solves
Information is split between SharePoint, OneDrive, Outlook and Teams
The hub connects through Microsoft Graph to SharePoint, OneDrive, Outlook and Teams data within approved scopes. Staff ask from Microsoft Teams, and each answer links to its source.
Staff leave Teams to search for documents in other places
People ask the hub inside Microsoft Teams, where they already work. Answers arrive with a link to the source, so the document is one click away.
IT must control exactly what an AI assistant can touch
IT controls every Microsoft Graph permission the hub uses. The field mapping is documented, up to three permission groups apply, and testing runs on sandbox or test records.
A good fit if
Policies, procedures and project files sit in SharePoint sites and OneDrive folders, and people still ask a colleague where the current version is.
Useful answers are buried in Outlook mail and Teams channels, and you want them reachable within scopes IT approves.
Your IT team can create an Azure app registration with Microsoft Graph permissions and give a 30-minute setup call.
Your staff already work in Microsoft Teams all day, so asking a question there beats opening another tool.
Not the right fit if
Your company runs on Google Drive, Docs and Gmail. See AI Company Knowledge Hub for Google Workspace.
Your team wiki is in Confluence rather than SharePoint. See AI Company Knowledge Hub for Confluence.
You mostly want decisions and lessons captured from meetings and threads into searchable memory. See AI Company Memory.
Your questions are mainly HR and compliance policy, answered with the exact clause. See AI Policy Q&A.
How it works
- Your IT team creates an Azure app registration with the Microsoft Graph permissions agreed on a 30-minute setup call with your Microsoft 365 admin.
- We document the field mapping: which SharePoint sites, OneDrive folders, Outlook mailboxes and Teams channels are read, what may be updated, and which permission group each belongs to.
- Up to 2,000 documents from Microsoft 365, plus up to 1 more system, are indexed for retrieval with citations.
- Staff ask in Microsoft Teams, a web app or Slack, and each answer quotes the passage and cites the SharePoint page, file, email or Teams message behind it.
- Up to 3 permission groups, configured for your Microsoft 365 setup, keep restricted sites and mailboxes away from the wrong people.
- 60 benchmark questions run against live sandbox or test records before launch, and a monthly refresh picks up new and changed content.
What you provide within 72 hours
Your portal lists these the moment you order. Delivery starts the working day after everything is in and your access is checked.
- An Azure app registration with the Microsoft Graph permissions agreed for the build.
- A Microsoft 365 admin for a 30-minute setup call to confirm scopes and the field mapping.
- The SharePoint sites, OneDrive folders, Outlook mailboxes and Teams channels to include, and those to leave out.
- Your user list and the permission groups they belong to.
- 30 sample questions and the answers you expect, taken from real Microsoft 365 content.
- Access to one additional system if you want it, and a single approver for decisions.
- An API key for your own AI model account (for example OpenAI or Google Gemini), created for this build with a spending limit.
The 72-hour rule. Send everything on your checklist, including access, within 72 hours of ordering so your build starts on time.
Timeline
Start
App registration in place, setup call done, Graph permissions and mapping agreed.
Weeks 1–2
SharePoint, OneDrive, Outlook and Teams content in scope indexed and grouped.
Week 3
Additional system and the Teams, Slack or web app interface connected.
Weeks 3–4
60 benchmark questions on sandbox or test records, accuracy report sent.
Week 4
Your review and two rounds of changes.
Go-live
Live, recorded handoff, field mapping document handed over, 30 days of support.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $12,400 |
| Express | 15–25 working days | $13,650 |
| Rush | 10–20 working days | $14,900 |
What you receive at handover
- The AI company knowledge hub for Microsoft 365, available in Microsoft Teams or a web app
- A two-way Microsoft 365 connection through Microsoft Graph, with documented field mapping
- Cited Q&A over up to two thousand documents, with three permission groups and a monthly refresh
- A test report on sandbox or test records, with sixty benchmark questions and an accuracy report
- An admin guide and the recorded 60-minute handoff session
- 30 days of post-launch support
Expected benefits and how to measure them
- Can give staff answers without leaving Microsoft Teams
- Helps IT approve the hub through permissions it already manages
- Makes SharePoint content easier to use
- Document search time
- Time a set of real searches across SharePoint and OneDrive before launch. Repeat them with the hub.
- Hub questions asked in Teams
- Track weekly question volume from the hub's usage records.
- Benchmark accuracy
- Re-run the sixty benchmark questions after a few months and compare with the launch report.
Before you start: List the SharePoint sites and mailboxes in scope. Time a set of real searches and note how often staff ask IT where files are.
These are things to track, not promised results. Outcomes depend on your data, your team and your market.
Running costs
You pay for the build once. Monthly running sits in your own accounts at roughly $50 to $300 for AI model usage, the vector database and hosting. Microsoft 365 licenses stay with Microsoft as they are, and plan upgrades are outside this build.
AI usage during build and testing is billed to your accounts as well.
Ask about running costsCustom requirements
Custom Microsoft 365 objects beyond the agreed mapping, an extra system, or more than 2,000 documents can be scoped through a note at checkout. Expect a written Change Quote; that work begins only after you accept it. Larger or bespoke projects go through the Custom AI request form.
Request a custom versionKeep it running
Monthly maintenance covers monitoring, fixes, model and platform updates, a monthly health report and monthly hours for changes, such as adding a SharePoint site to scope or adjusting Graph permissions with your admin. It starts after the 30 days of support.
$620/mo (5% of the build price a month, includes 6 change hours)
Optional. Starts after the post-launch support period. Cancel any month.
How maintenance worksHow this compares
| Do it yourself with a SaaS tool | Hire a custom agency | Upmark Labs | |
|---|---|---|---|
| Price | A per-license AI add-on bills for every user each month | An IT partner quotes after a tenant review | Ours is one fixed price, shown before you buy |
| Permissions | Follows whatever sharing users have set | Depends on the design agreed | Ours uses agreed Graph permissions plus up to 3 permission groups |
| Where staff ask | Usually tied to one app | Built to order | Ours runs in Microsoft Teams, Slack or a web app |
| Delivery | Switched on fast, then weeks of clean-up | Usually several weeks | Ours has a fixed delivery time shown before you buy, with faster options |
Questions about this product
What Microsoft 365 data does the knowledge hub read?
SharePoint, OneDrive, Outlook and Teams data, within the Microsoft Graph permissions your admin approves. On the setup call you choose the sites, folders, mailboxes and channels in scope, and the field mapping records each one. Content outside those scopes is not read, and custom Microsoft 365 objects beyond the mapping are excluded.
What does our IT team need to set up for the Microsoft 365 edition?
An Azure app registration with the Microsoft Graph permissions agreed for this build, plus a Microsoft 365 admin for a 30-minute setup call. Those permissions decide what the hub can reach, so they are confirmed before indexing starts. Arranging the registration within the 72-hour intake window keeps delivery on its date.
Can staff ask the hub questions inside Microsoft Teams?
Yes. Microsoft Teams is one of the three interface options, alongside a web app or Slack, for up to 50 users. Staff ask in a chat, and answers come back with a citation to the SharePoint page, OneDrive file, email or Teams message behind them, so anyone can open the source.
Will the hub edit our SharePoint or Outlook content?
Only inside the agreed field mapping. The Microsoft 365 connection is two-way, but every update is limited to the scopes and fields documented and approved at intake. Its main work is answering questions with citations, and the mapping document you receive at handoff shows exactly what it can touch.
When can our staff start using the Microsoft 365 hub?
Delivery runs on a fixed schedule from a verified intake, and the delivery panel shows that date plus express and rush options before you order. The clock starts once the Azure app registration, the setup call and your sample questions are complete, so involve IT on the day you order to keep the date.
What does the Microsoft 365 hub cost per month to run?
Roughly $50 to $300 a month in your own accounts for AI model usage, the vector database and hosting. Microsoft 365 licenses continue as they are and are not part of this build. Monthly maintenance is optional, if you want changes made after launch.
Can we connect a non-Microsoft system as well?
Yes, up to 1 more system is included, such as a CRM or a Confluence space, indexed and cited alongside your Microsoft 365 content. A further system comes through the Knowledge Base Expansion add-on, or a Change Quote agreed before that work starts.
Let Microsoft 365 answer back, source attached.
Order AI Company Knowledge Hub for Microsoft 365 and staff get cited answers from SharePoint, OneDrive, Outlook and Teams on a fixed delivery date, inside the permissions IT approves.
Works well with
Knowledge Assistants
AI Policy Q&A
Answers employee questions about policies with the exact source clause.
