CampusMind Data Bridge · Built on Microsoft Fabric

Build the data layer campus AI needs.

Data Bridge connects your campus systems in Microsoft Fabric once. Every AI use case after that starts from shared records and governed access, not from scratch.

Get a free data assessment Try the Data Bridge lab

Your systems stay where they are. Your data stays in your tenant.

A question goes all the way down. Only the answer comes back.

PeopleStudents, faculty, advisors, staff

CampusMindAdvising, registration, planning

Data BridgeOntology, definitions, access rules

Microsoft FabricOneLake. You own this.

Source systemsSIS, LMS, CRM. Nothing replaced.

01 · The problem

Every new AI use case starts from scratch.

Most campuses have run an AI pilot. The assistant handles the course catalog well. Then someone asks which advisees are falling behind, and it has nothing useful to say.

That isn't a model problem. A student means one thing in the SIS, another in the LMS and something else again in advising. So every new use case rebuilds the same connections. It happens at every institution.

Today · point to point4 systems × 8 use cases

32separate paths. Each one maps the records, defines the terms and sets up access again.

With Data Bridge · connect oncesame inputs

12paths. Records mapped, terms defined and access governed one time, then reused.

Connect once. Reuse everywhere.

02 · What Data Bridge builds

One governed layer between your systems and every agent.

Your systems keep running as they are. Their data lands in one governed place in Microsoft Fabric, and Data Bridge adds the meaning on top.

Where it goesEnrollment · Advising · Academic operations · Finance and registration · Institutional research · Executive planning

CampusMind Data BridgeFabric IQ in preview

01A shared vocabularyPeople, programs, courses, terms and how they relate, written down as a model instead of living in people's heads.Ontology

02One definition of every numberRetention, at-risk and credit completion, defined once by the people who own them and applied everywhere.Semantic models

03Only the fields a task needsEach agent gets the approved, task-specific fields for its job and nothing more.Governed data provisioning

04Answers that cross systemsAgents draw context from the whole governed layer, so one question can span enrollment, holds and advising.Agent orchestration

Microsoft Fabric OneLakeraw → cleaned → certified · full lineage · you own this

Your campus systemsSIS · LMS · CRM · Finance/ERP · Advising · nothing replaced

Isolates critical systemsUnifies governanceLimits data accessPreserves traceability

03 · Why it answers differently

Ask one system. Then ask the Data Bridge.

Synthetic demonstration data

Ask the source system“How many event registrations do we have?”

1,200rows from one table. It can't tell a student from a staff member.

Ask the Data Bridge“How many students actually attended?”

398students. Engagement platform plus SIS, read through one definition of a student.

Students who attendedStaff, faculty and no-shows1 dot = 1 registration

04 · Answers you can check

Every answer shows how it got there.

Ask in plain language. The answer comes back with the path it took through your data, so your team can check it before anyone acts on it.

Product UI · synthetic data

Which graduating seniors are blocked from registering?

Fabric Insights Agent

Hold type	Students	Blocks registration
Financial	312	Yes
Advising	406	Yes
Immunization	58	Yes

How I got this

Ontology path: Student → Enrollment [Fall 2026] → Holds

Measure: Hold count

Filters: graduating seniors, hold type = registration block

Source tables: fact_lms_weekly, fact_hold

Refresh time: Sep 19, 2026

Can you create a report of this?

Your definition of a hold

Counts come from certified definitions. Agents can't invent a metric.

The path it walked

Every entity and relationship the answer relied on.

What it read, and when

Source tables and refresh time. Read-only, always.

Turn any answer into a report your office can share.

05 · Security and governance

The AI doesn't decide who sees what. Your permissions do.

Every question runs under the asker's own Microsoft Entra identity. Access rules sit in the data layer, underneath the AI, so no phrasing of a question reaches a row that person isn't entitled to.

Your identity, not a service account

Every query runs as the person asking, through Microsoft Entra.

Row and column security

A boundary in the data, not an instruction the model could ignore.

Read only, with a full audit trail

Agents can't change records. Every question and answer is logged in Microsoft Purview.

Security reviewer? Request the technical architecture brief →

Same question. Same layer. Different rows.Try a role

Student1

Advisor118

Provost1,000

1of 1,000 student records reachable

A student sees their own record. Nothing else exists for them.

118of 1,000 student records reachable

An advisor sees their caseload, not the institution.

1,000of 1,000 student records reachable

The provost sees the full cohort. Still read only.

Access to student education records follows your institution's own authorization rules, with a record of who asked what.

06 · What becomes possible

One foundation. Every office that needs an answer.

Once the layer exists, each new use case builds on it instead of starting over. These are real questions from the Data Bridge lab.

Leadership“Which program has the most students at risk?”“Which advisors are supporting those students?”

Advising · caseload only“Show me a few students struggling with missing assignments.”“What's going on with this student?”

Registrar“Which graduating seniors are blocked from registering?”

Student engagement“How many students attended our events this term?”

Students, through CampusMind“Do I have anything that could stop me from registering?”“What financial aid do I have on file for Fall 2026?”

Your next oneBring your first cross-system question.We'll scope it in the free assessment →

Try these questions in the lab

07 · Ownership

Built on Microsoft Fabric. Owned by you.

No new vendor cloud to review. Data Bridge lands inside the Microsoft estate your team already operates.

Fabric capacity is often fundable through your existing Microsoft agreement.

01
Your identity system
People sign in with the Microsoft Entra account they already use. No second directory to manage.
02
Your governance tooling
Microsoft Purview governs this the way it governs your Microsoft 365 data.
03
Your tenant, open format
Data sits in open Delta format in your own tenant. End the engagement and you keep a working platform.
04
Your definitions
Institutional Research owns every metric. Data Bridge makes those definitions binding everywhere, including for the AI.

08 · How we start

Start with one real workflow.

The free data assessment reviews your systems and scopes one cross-system use case with you. The findings are yours either way.

AssessAudit the data and BI estate you already have, and name the gaps worth fixing.
DesignSet the target architecture, the governance model and its security boundary.
BuildMigrate the data, rebuild the pipelines, and tune the reports people actually open.
ScaleTrain your teams, strengthen the models, and assign agents to data.

Get a free data assessment

Questions IT asks first

Straight answers.

We already have a data warehouse.

Good. That's a source, and it shortens the first phase. What a warehouse doesn't give you is a layer that tells an AI what your numbers mean and what each person may see.

We just finished a Banner or Workday implementation.

Then you're well placed. Those are sources, not obstacles. Nothing gets replaced, and the work sits on top.

Is Fabric mature enough to bet on?

The foundation is generally available. Parts of Fabric IQ are still in preview, and we'll tell you which ones and what our fallback is for each.

Why Fabric and not Snowflake or Databricks?

If you already run Microsoft 365 and Entra, Fabric keeps you on the identity and governance tooling your security team has already reviewed. The others each add another review cycle.