SharePoint, Google Drive, Nextcloud and network shares stay where they are. WebChatAgent indexes your folders, answers questions from them and can show the exact file location.
Before / after
Ask your team where the current travel expense policy lives. You will get three answers, two of them pointing at outdated versions. The reason is rarely laziness, it is history: every department filed things wherever it made sense at the time, and over the years four systems grew six different folder logics. Cleaning up fails because nobody has the time and everybody is afraid of deleting something important. WebChatAgent takes the other route: the bot connects to the folders as they are, reads the files and answers questions from them. Anyone who wants to know how a process works asks the bot instead of the colleague who has answered that same question three times already.
Is that in Teams or in Drive now?
Three colleagues are asked, two versions are opened and nobody is certain.
Where can I find the current travel expense policy?
The current policy has applied since January. Here is the approved version:
Last synchronized: today, 08:14
Features
Enter the credentials for your system and pick a folder. Everything readable inside it goes into the index. New files are picked up on their own.
You set an interval, daily to monthly. Each run reads only what changed. Deleted files disappear from the answers.
On request the bot names the folder path alongside its answer. Anyone who needs the file itself knows where to go. For public widgets this stays off.
Scanned PDFs go through text recognition, spreadsheets keep their row structure. Google Docs and Sheets are converted automatically.
Exclusion patterns like /Archive/** or *_internal* apply while reading. Individual files are removed from the index with one click, for good.
The connection needs read access and nothing more. Credentials are stored encrypted, processing runs in a German data centre. Revoke the share and it stops.
In daily work
In the first weeks every other question is a search question. The bot answers them around the clock without interrupting anyone.
Instead of hunting through three systems for the right warranty clause, the team asks the bot and gets the passage with its source.
Which terms apply at this volume? The bot answers from the price list updated yesterday, not the 2024 PDF buried in someone’s inbox.
Maintenance instructions, safety data sheets, plant documentation. Everything stays where it belongs and is still one question away.
Vacation policy, travel expenses, parental leave: the same question reaches the HR inbox for the twentieth time. The bot answers from the current version, not the PDF from two years ago.
In an audit, the evidence matters as much as the answer. The bot cites the applicable procedure and tells you which folder it is stored in.
On-site with a customer, without VPN access and no time to spare: plant documentation and spare parts lists are one chat question away.
When someone leaves the team, their knowledge remains in their folders. Once indexed, it stays searchable without needing to ask that person.
Decision guide
Compare document sources, Office integration, employee permissions, channels and the different billing models.
No, and that is the entire point. The bot works with the structure that grew over time and walks subfolders however deeply they are nested. The only thing worth doing up front is excluding the areas nobody outside a specific team should see.
SharePoint and OneDrive through Microsoft Graph, Google Drive through a service account, Nextcloud and ownCloud over WebDAV, and your own servers over SFTP or FTPS. Unencrypted FTP is deliberately not supported. Confluence and Notion are available as separate connectors.
You choose the interval, from daily to monthly. Each run checks which files changed and reads only those again. That keeps cost down and answers fresh. You can trigger an immediate refresh by hand at any time.
It sees exactly what you grant access to. Within that, patterns exclude whole areas, for example /HR/** or anything with _confidential in the name. Individual files are removed in the overview with one click and do not come back on the next sync.
They are stored encrypted and only sent to the connected system to read files there. Once saved, the dashboard no longer shows them. The connection needs read access only, the bot cannot write anything.
PDF including scanned documents, Word, Excel, CSV, plain text and Markdown. Google Docs, Sheets and Slides are converted automatically on retrieval. Images and videos are skipped.
Step-by-step tutorial
Follow the detailed video and nine illustrated steps to connect a read-only folder, exclude restricted paths and prove a policy update from USD 68 to USD 72 end to end.
Related guides and pages that pair well with this one.
Turn your Confluence wiki into an instant AI answer bot.
Read moreAnswer questions straight from your Notion workspace.
Read moreHow teams turn Notion, Confluence and internal docs into one AI answer layer.
Read moreFive transparent plans, from free to enterprise.
Read moreEnter the credentials, choose a folder and start the first indexing run.
