# FAQ and troubleshooting

Your PHOENIQS account signed in, but it is not among the people your organisation admits to this Data Pool. Ask an administrator of your company for access.

Data Pools needs the PHOENIQS AI Model Service to index, read and describe documents. While it is not configured or not reachable, uploads, OCR and descriptions fail. Contact support.

Large files and scans take longer, because every scanned page is read by an AI model, and files are processed a few at a time. If a document does not finish, delete it and upload it again, and contact support if it stalls again.

Open the document's Details for the reason. Then delete the document and upload it again. If it fails again, contact support with the reason shown.

Its OCR reading may be incomplete, for example because of a poor scan, handwriting or an unusual layout. Agents are told the same. Open Details to see what was read, and upload a better scan if you can.

Documents can be up to 1 GB, CSV files up to 25 MB. Uploads also stop once your Data Pool's storage is full: the message reads "This datapool's storage is full" and shows how much is used. Delete documents or tables you no longer need, or add Data Pool Additional Storage.

The app does not show it yet. Contact support to find out. Once the storage is full, the upload message shows how much is used. To get more storage, add Data Pool Additional Storage in the PHOENIQS Portal.

Files in a SharePoint pool mirror the SharePoint site. Delete the file in SharePoint, and it disappears from the pool at the next sync, within 30 minutes.

• Check the document is Ready in its pool.
• Ask an administrator to try the same question in the pool's Test retrieval. If the passage comes back there, the pool is fine and the agent may be searching another pool: a clear pool description helps it choose.
• Files uploaded in Chat are in your AI Chat pool, which agents reach through Chat's file search, not through the MCP server. See Chat and agents.

No. The MCP server only offers tools that list, search and read. Uploads, deletions and access changes happen in the Data Pools app.

Only pools you have added them to as a member, and pools open to your organisation. An agent always acts as its user, and sees exactly the pools that user can read.

In the pool's Access section, add them by e-mail address. To open a pool to everyone in your organisation, switch it to Organisation. See Access.

They have not signed in to Data Pools since you added them. Their access starts at their first sign-in.

Uploaded files, tables and their indexes are stored by PHOENIQS in Switzerland. SharePoint files are not copied: only their details and links are kept. External databases are queried where they are, read-only.

No. Indexing, OCR and descriptions run on the PHOENIQS AI Model Service, which does not use your data to train models.

It is removed together with everything derived from it, such as its passages, description and metadata, at once. It cannot be undone.

Contact support.