Cori, the team's AI assistant, and evaluation analyses

What changed

In June 2026, the clinic team got an assistant, Cori, and evaluations gained two things: AI-generated analyses and a comparison between evaluations.

Cori, for the team

Cori opens from a button in the corner of the screen, anywhere in the app, and replies in the language you are working in. It is for therapists, coordinators and administrators only; parents cannot access it. Cori is included in every plan except Starter. What it does:

• Answers questions about the app: how to book a recurring session, how to compare two evaluations, how a programme is scored.
• Looks up a child's data: evaluations, recent sessions, the intervention plan and goals. It only gets billing data when an administrator asks for it.
• Helps with interpretation: explains scores and session history, suggests where therapy might be worth focusing, and can draft a note or a goal.

Cori supports clinical judgement rather than replacing it. It gives no diagnoses, and its suggestions are options to discuss as a team. Conversations stay in each user's own history.

What data Cori receives

Before their first question, each team member consents to their data being sent to Cori's AI service. Without this consent, Cori receives nothing.

When Cori reads a child's record, the names of the child and the therapists reach it only as initials, age instead of date of birth, with no phone or email address. The child is identified by a code that contains nothing from their name. The full list of data is in the Data Processing Agreement, section 7.

What you write in the conversation reaches Cori exactly as written. So there too, use the child's initials, not their full name.

AI analyses of evaluations

For every completed evaluation, you can generate an analysis: a summary, strengths, where to focus, and notes by category. The scores, the child's initials, their age and the diagnosis are sent for the analysis, without the child's name or the family's contact details. It reads the results according to age; for one evaluation, sections not yet expected at the child's age are not treated as gaps.

The analysis is kept with the evaluation, with the date and the name of the person who generated it, and can be redone. It carries a note that it is AI-generated and must be checked before clinical use.

Comparing evaluations

When a child has at least two completed evaluations of the same type, the "Compare evaluations" button puts them side by side. You choose which two, and the app shows the overall change and, by category, what has improved, what has stayed the same and what is worth watching. The comparison works for every evaluation in the app.