Skip to content

CeliaConnect User Manual

Every screen in CeliaConnect, explained.

One page per customer-app surface. What it is. Who uses it. What you can do. Common workflows. Common gotchas. Built for the day-to-day work — read what you need, ignore the rest.

28 surfaces documented. Updated as the customer-app evolves — see the sync-check script for the drift contract.

What is CeliaConnect?

An AI assistant that reads every record overnight — without ever seeing a name.

CeliaConnect is an AI assistant that reads every prospective student’s record overnight — without ever seeing names, emails, phone numbers, addresses, or any other personally identifying information — and produces five outputs per student: Engagement (how active they are in the funnel right now), Readiness (how close their application file is to a decision), Yield (how likely they are to enroll if admitted), Risk (a low / medium / high / critical tier that combines the three), and a recommended next action (the single most useful thing a counselor should do next, with a one-line reason).

The AI sees only anonymized signals from your institution’s own data — a strict architectural guarantee, not a policy promise — so every analysis stays inside your tenant boundary and FERPA-aligned by design.

A note on PII

Every example here uses anonymous IDs.

CeliaConnect's no-PII architecture is structural, not aspirational. Every "imagine a student row" example in this manual uses anonymized identifiers like s_8f42a91c — the same shape Celia actually sees. Your real students are not in here, and they are not in any prompt Celia ever sees.