How it fits with campus governance

Both evaluate and decide, at different moments. The AI Strategic Compass decides which initiatives to start and how closely to review them. This framework decides whether they work once running, and what’s next.

Before you commit

Is it worth doing, and how closely should we look?

Once it runs

Is it working, for whom, and at what cost?

The relationship at a glance

The Compass decides what gets started. The Campus AI Registry holds the record. This framework judges what happens next, and sends its decisions back.

Before commitmentAI Strategic Compass

Is it worth doing, and how closely should we review it?

ScreenScoreSelectPlan

OutputProceed, pause, or revise; review band; success measures

Selected initiative, review band, success measures

The shared recordCampus AI Registry

Who owns it, what data it uses, and what it should achieve?

OwnerWorkflowData classRiskBaseline

OutputA registered use case with a baseline

Registered use case with a baseline

Once it runsThis framework

Is it working, for whom, and at what cost?

MeasureEvaluateDecide

OutputContinue, improve, expand, restrict, or retire

Every decision updates the registry. This framework → registry.

A major change triggers re-review. New data, a wider audience, or more autonomy: this framework → Compass.

Campus AI governance Strategy, principles, policies, risk appetite, accountability.

Not every use case starts at the Compass

The Compass is for initiatives needing a commitment. Everyday use of approved tools skips it. Every route ends in the registry and this framework, so all use is judged the same way.

Route A · Full review New initiatives New tools, purchases, integrations, or any use that meets a trigger below.

CompassRegistryThis framework

Path: Compass review, then registration, then evaluation. Example: The first-year advising assistant, which is new, student-facing, and uses student records.

Route B · Already approved Individual and team use Using an approved campus tool, within its approved data rules, for one person’s tasks or a team workflow.

Light registrationThis framework

Path: A short self-registration, then evaluation at the Individual or Team level. No Compass review. Example: Staff drafting meeting summaries in the campus AI assistant.

Route C · Found in use Unapproved or unregistered use Personal or unapproved tools found through surveys, spending reviews, or self-reports.

DiscoveredRegistryCompass

Path: Register what is found, then send it to the Compass if a trigger applies. Otherwise, move it to an approved tool. Example: A department paying for a personal chatbot subscription to answer student email.

Send it to the Compass if any of these apply

A new tool, vendor, or contract
Confidential or regulated data (for example FERPA or HIPAA)
Student-facing, or affects grades, admissions, aid, or employment
Delegated actions, where AI acts without a person approving each step
Use spreads beyond one unit, or needs new funding

What light registration records

Tool, task, data used, and who reviews the output. It takes a few minutes and goes into the same registry as full reviews.

Individual uses are counted together by tool and role, not evaluated one by one. They show up in the use funnel and in the institution-level view of the campus platform.

If an individual or team use grows or meets a trigger, it moves to Route A.

Side by side

	AI Strategic Compass	This framework
Main question	Should we start this initiative, and how closely should we review it?	Is this use case delivering value at acceptable cost and risk, and for whom?
When	Before commitment: at proposal and selection	During and after the pilot, then on a review cycle
Unit of analysis	An initiative or proposal	A use case in a defined workflow
How it scores	Five dimensions rated 1 to 5: strategy, ethics and compliance, finances, operations, stakeholder impact	Cost, Use, and Outcomes measured; five lenses checked as pass, watch, or fail
Proportionality	Four review bands, from low to critical	The same four bands set the quality, evidence, and approval bar at evaluation
Output	Proceed, pause, or revise; success measures defined	Continue, improve, expand, restrict or pause, or retire
Levels	Initiative scope	Individual, team, unit, institution

Feeding existing review cycles

From the use case inventory

This framework’s decisions go into reviews the campus already holds, using the same scorecard. No new report.

Existing cycle	When	What it receives
Budget planning	Annually	Portfolio costs, Continue and Expand requests, retirements that free funds
IT portfolio review	Quarterly or annually	Platform and vendor decisions, lifecycle and exit risk, duplicate tools
Academic program review	On the program cycle	Teaching and learning use cases, equity results, student outcomes
Risk and compliance review	Annually, plus incidents	High and Critical band use cases, incidents, data and accessibility findings

Decision words, matched

The two use different words for similar choices. Pick one set for campus communication.

Compass	This framework	Meaning
Proceed	Continue	Results meet pilot requirements.
Iterate	Improve	Value is plausible; a specific weakness needs correction.
Scale	Expand	Evidence, safeguards, and capacity support broader use.
Pause	Restrict or pause	Quality, harm, access, or control requirements not met.
Stop	Retire	Insufficient value, or a better alternative exists.

Worth reconciling

Two scales

The Compass rates dimensions 1 to 5. The inventory scores here run 0 to 1. Say plainly which applies before launch and which after.

Equity, checked twice

The Compass checks equity at intake; this framework checks it in results. Treat them as one continuous check.

Who owns the handoff

Name who moves a use case from selection to registration, so success measures do not get lost between offices.

The AI Strategic Compass and the Campus AI Registry are separate tools. Descriptions here summarize their public sites; see campusaicompass.com and campusairegistry.com for details.