Frequently asked questions

With the LMS your institution already uses, through its web services and in read-only mode. Before any pilot we check together that your platform has those services enabled. Today Pharis connects to Moodle; the details are in Early warning in Moodle.

No. Nothing is installed on the LMS server and no administrator access is needed. Professors and administrators use a web browser; students use a Chrome extension or a lightweight Windows app.

No. Pharis explains the method using their course material and guides them to reach the answer themselves. Students always know they are talking to an AI.

It uses only what the student does in the LMS: days with access compared with their group, missing or late assignments, and grades. It is trained on public data from other universities, so we do not publish an accuracy figure: it is calibrated at the end of a full term, with your institution's actual dropouts. Every alert shows its reasons. Details in How Pharis calculates dropout risk.

No. The risk level is a signal for a person, who decides whether and how to act. The signal suggests support, never sanctions.

Yes, optionally. If the institution shares signals such as an outstanding balance or a scholarship, they are added to the risk with a fixed, explained adjustment, and the finance team sees only what concerns them.

Pharis only reads from the LMS, everyone sees only their own data, and we do not train AI models on your data. When the contract ends we delete everything, and backups expire within 30 days.

With a demo on your institution's real data: we connect your LMS in read-only mode and show you your groups, the students at risk today and why. Then an 8-week pilot with a few volunteer professors. If you do not continue, we delete the institution's data.