Institutional research, governed
Answers backed by your definitions.
InstitutionalResearch.AI changes how your campus works with its data. Every metric runs on one definition verified by your experts and tailored to your needs, so each question has one repeatable answer. Streamline reporting from census to IPEDS: pick a ready-made report or ask in plain English, and get the same answer every time, with its definition and scope attached.
Now seeking additional partner institutions. Built by M-Powered Insights.
Why it matters
Every office has its own retention rate
Institutional research spends its year reconciling numbers, rebuilding the same reports, and deciding who may see student data. Generic BI tools and AI chatbots don't solve any of that.
Numbers that don't match
Admissions, the registrar and the provost's office each count "enrolled" differently, and the board sees three versions. A number nobody can trace can't be defended.
IPEDS season, every season
Each survey is rebuilt by hand from extracts and spreadsheets, then checked against last year's submission, under a federal deadline.
Access left to good intentions
FERPA, small-cell privacy and "advisors see only their advisees" usually live in policy documents, not in the system that answers the question.
How it works
From your student system to an answer you can defend
-
Connect
We map your student information system into our standard IR data model. Your experts review how every field is mapped, and every data load comes with a data-quality report.
-
Define
Every metric has one definition and one calculation, and every change to them is logged. Campus choices, like gateway courses or credit thresholds, are settings you choose, not custom code.
-
Ask
Pick a ready-made report or ask your question in plain English. The governed engine computes the answer within your role and scope, and shows the definition and time period behind every number. Adjust your question on the fly, with no need to retype.
-
Report
Download PDFs, schedule reports delivered as secure links, and build IPEDS upload files that reconcile with your governed numbers before you can download them.
What's included
Routine reporting, handled. Time back for the work that matters.
29 governed metrics and growing
Census enrollment, admissions funnel, retention and persistence, graduation rates, completions, credit momentum, course outcomes, faculty workload, aid and more. Each one is checked against an independent calculation.
IPEDS upload files
Fall Enrollment, 12-month Enrollment, Completions, Graduation Rates, Outcome Measures, Admissions and Human Resources, built from the same governed metrics, laid out like the survey forms for review, and compared with what you submitted last year.
Peer benchmarks
Compare retention, graduation, admissions, salaries, Pell share and net price with a peer group you choose, using NCES's published values, always for the same IPEDS year.
Student lists, within scope
Open the students behind a number to act on it: an advisor sees only their advisees, a chair only their department. Small groups are suppressed automatically.
Reports and schedules
Census and end-of-term reports, any answer as a PDF, and schedules that email a link, never an attachment, so access is checked every time a report is opened.
Fits the data you have
If your institution doesn't provide something, like salaries or test scores, the features that need it are hidden rather than showing a wrong number.
Not another AI chatbot
The AI understands the question. It never decides the answer.
In most AI analytics tools, a language model writes the query and summarizes the result, so answers drift and access rules live in a prompt. Here the AI does one thing: it turns your question into a request for a governed metric.
Correctness, access, scope and privacy are enforced by the engine and the data layer. Switch the AI off and every one of those protections still holds; you simply pick the metric yourself.
What the AI sees
- Your question, and the request you're following up on
- The catalog of governed metrics and their definitions
- Your institution's lists of programs, colleges and departments, so it can filter by them
What it never sees
- Student or staff records, or results: answers are computed after the AI is done
- Access rules: they're enforced in the data layer, not in a prompt
Who it's for
For the people who report, decide and protect
IR directors
Definitions your office signs off on, IPEDS files that reconcile, and fewer ad hoc requests, because deans and chairs can answer their own questions within their scope. More time for the analysis only your office can do.
Provosts, deans and chairs
Retention, enrollment and course outcomes for your college or department, the same numbers IR reports, with peer comparisons for program review.
CIOs and security offices
Single sign-on through your identity provider, role and row-level access enforced in the data layer, a complete audit trail, and an AI that never sees student records, so none can be used to train it. Trust & security
Get started
Become a partner institution
We work closely with each partner to load your data, verify the numbers against what you've already published, and shape the product around your IR needs. If that interests you, we'd like to talk.
Try the demo, or talk with us!
The demo runs on a fictional university and lets you try the product from each role, from advisor to IR director. Test-drive everything from IPEDS files to peer comparisons at the click of a button. Request access with your college or university email, and we'll send you a 14-day access code!
Or email hello@institutionalresearch.ai.