AI agents for SharePoint document storage

Covera Agents reads the agency's SharePoint documents, read-only: files are attached to a quoting or comparison file without downloading them again, and the assistant searches the folders the agency allows.

Covera Agents builds AI agents for insurance brokers. In many agencies, policy documents, questionnaires and statements live in SharePoint, on Microsoft 365. The SharePoint connection lets the agents use those files where they already are, without anyone downloading and uploading them again.

Read-only by design

The permissions Covera asks Microsoft for are read-only. Even an error on Covera's side could not modify or delete a document. SharePoint is not a portfolio connection: it holds files, not clients or policies, so nothing from SharePoint feeds the agency's book in Covera, and Covera never guesses a client from a folder name.

The folders the agency allows

After connecting, an agency administrator ticks the folders Covera may display and use. The scope is closed by default: until a folder is ticked, Covera shows no file and does not even query Microsoft. Allowing a folder allows everything it contains, including subfolders created later. Each broker then sees the allowed folders within their own Microsoft rights, and the check is done again at download time.

What it unlocks for the agents

In the import step of a carrier quoting agent or quote comparison agent file, the broker browses sites, libraries and folders, or searches, previews a file and ticks the ones needed. Covera downloads them on the server side and attaches them to the file. If a download fails, the request fails, rather than starting an analysis on fewer documents than ticked. The assistant can also search the allowed folders and read a document, for example to find a clause in a policy.

How the connection works

Each broker connects their own Microsoft account from Covera's settings. SharePoint uses a separate Microsoft app from Outlook, so it has its own consent: a Microsoft 365 administrator often approves it once for the agency, then brokers connect on their own. No site-by-site setup is needed.

Traceability

Every read is logged: who, which agency, which file.

Browsing large libraries

A renamed folder stays allowed, and a folder name reused elsewhere opens nothing. Very long lists are shortened in the picker: go down one level or search instead. One import handles up to about 500 MB of files.

Limits

Covera does not write to SharePoint, so received documents are not filed there automatically. An entire library cannot be ticked at once: its top-level folders are ticked instead. Personal OneDrive storage is not covered. Covera does not train any AI model on agency data.

Other integrations

Mailboxes connect through Outlook, with its own Microsoft consent, or through Gmail.

Frequently asked questions

Can Covera Agents modify or delete our SharePoint documents?

No. The Microsoft permissions requested by Covera Agents are read-only, so even an error could not write to SharePoint.

Does Covera Agents see our whole SharePoint?

The Microsoft permission covers what the connected account can see. Covera Agents only displays and uses the folders an agency administrator ticks; with nothing ticked, nothing is shown.

Can the Covera Agents assistant find a clause in a SharePoint document?

Yes. It searches the allowed folders, then reads the document to find the clause.

Other integrations

Meet the Covera team

We show you Covera at work: how the agents handle your agency's servicing requests, from the incoming email to the updated policy file.