Skip to content
Campus AI Framework Campus AI Match
Find your AI tool
Concept demo. Tools, approvals and contacts are examples, not any campus’s real policy. Catalog last updated 23 Oct 2026 · Rules v1.0
A concept from the Campus AI Framework

What if finding the right AI tool on campus took a minute?

Campus AI Match is a working concept. It shows how a campus could guide people to approved AI tools, or to the right person, using decisions the institution has already made. The tools and rules here are examples.

Why this matters

Colleges and universities now offer a growing range of AI tools and services. Each serves different needs across students, faculty, researchers and staff, and each must also meet the institution’s obligations: data privacy and security, equitable access, accessibility, academic integrity and responsible use of funds.

With many options and many rules, people often can’t tell which tool to use or what they can safely put into it. Campus AI Match is a concept for closing that gap, turning institutional decisions into clear guidance for the people doing the work.

How it works

Start with the simplest case. Most requests stop at the first check; only new projects reach the end.

You have an AI request
See how example requests travel the paths →
For institutions

How a campus could build this

The demo works because a few decisions were made first. Your campus would make its own. Three short pages cover what to decide, how to start, and how to learn from what people ask for.

Licensed CC BY 4.0: free to share and adapt with credit to Joe Sabado.

Next: {{ instNav.nextLabel }} →
For institutions

Six decisions

Make them once, write them down, and review them each term. Open any row to see what to think about.

1You decideWhich tools you offer, and what data each can handle People getA tool catalog with a plain badge on every tool, such as “Up to Confidential”What to think about +
  • Is it secure, and does it use campus sign-in?
  • Does the vendor train on what people type? How long is data kept, and where?
  • What is the most sensitive data it can handle?
  • What can the AI read, connect to and do on its own? Review the tool around the model, not just the model.
  • Contract: FERPA “school official” terms, a BAA or data agreement where needed, who owns outputs, copyright indemnity
  • Admin controls and logs: campus sign-in, turning features off, pinning model versions, audit logs and records requests
  • Accessibility (a current VPAT), tested quality for the task, and whether everyone can get access or only license holders
  • An exit plan, an incident contact, and the environmental footprint if the vendor discloses it
  • What does it cost, and can you support it?
Review a tool again when:
  • Its review date arrives
  • The vendor adds a major feature, such as agents or web access
  • It gets new permissions or connects to new data sources
  • The vendor changes the underlying model
  • The vendor changes how it handles data
  • Its use spreads to more people, more sensitive data, or more units
Owner: AI governance lead; recorded in the Campus AI RegistrySee the catalog →
2You decideWho gets which tools by default People getA starter kit for each role: two tools, sign-in links, no formWhat to think about +
  • Which tools are open to everyone?
  • Which need a license, and who pays?
  • Which need an approved research project?
  • What experience should people have before tools that act on their own?
Owner: IT with unit leadersSee the starter kits →
3You decideWhat AI is never used for People getThree stop signs, and clear “not with AI” answers that say what to do insteadWhat to think about +
  • Will AI ever judge or rank people? We recommend no.
  • Which data never goes into any AI tool?
  • Is a personal AI account ever OK for campus work? We recommend no.
Owner: AI governance councilSee the start page →
4You decideWho handles sensitive or high-stakes requests People getA Get help page: a named office, what to bring, how fast they replyWhat to think about +
  • Which office owns each situation: student records, health data, export control, personnel, legal, integrity, hiring and admissions, licenses?
  • What should people bring?
  • How fast will each office answer?
Owner: Each officeSee Get help →
5You decideHow new ideas get reviewed People getA Propose page with a review sized to the stakesWhat to think about +
  • Which review process to use: your own, or adapt the AI Strategic Compass used in this demo.
  • What counts as a new initiative: a bot, an AI feature, a deployed model, an automated decision?
  • Who decides at each level, and how long should it take?
Low stakes · unit lead · daysModerate · AI governance lead · ~2 weeksHigh · AI council · 4–6 weeksCritical · executive sponsors
Owner: AI governance council, through the campus’s proposal reviewSee Propose →
6You decideWhat training people need People getAI basics for newcomers, required before tools that act on their ownWhat to think about +
  • What does everyone need to know?
  • What extra training comes before agent tools?
Owner: Teaching and learning center with ITSee the FAQ →
Next: {{ instNav.nextLabel }} →
For institutions

Getting started

Six steps to launch at your campus.

  1. 01List your approved tools and what data each can handle.
  2. 02Set who gets which tools by role.
  3. 03Pick two default tools for each role group.
  4. 04Name an office, contact and response time for each hard case.
  5. 05Point “Propose something new” at your own intake form.
  6. 06Test with twenty real requests alongside those offices. Where reviewers disagree, sharpen a rule.

How you’ll know it works

  • People reach a tool fast: two clicks for a starter kit
  • Most requests are handled without a consult
  • Consults are answered within a few days
  • Fewer people use personal AI accounts for campus work (by survey)

Review the catalog, kits and office list each term. Count in aggregate only; never record what people type.

Licensed CC BY 4.0: free to share and adapt with credit to Joe Sabado.

Next: {{ instNav.nextLabel }} →
For institutions

What people ask for

A match helps one person. Counted together, matches show where the tools fit, where they don’t, and what to buy, retire or build next. This closes the loop between what people need and the six decisions.

Counted

Role group, task, data level, who will use it, the path, and the tool matched or “no tool”.

Never counted

Names, accounts, files, or anything a person types. Small groups are combined so no one can be identified.

Six signals, and the decision each one feeds

Examples are illustrative. Some “no” answers are the rules working as intended; separate those from real gaps before acting.

  1. 1Tasks that end with no approved toolWhere the catalog has a gap For exampleCoding interview transcripts that still contain participant names Feeds decision01 Which tools you offer
  2. 2License requestsWhether a paid tool should become a default For exampleFaculty asking for licensed Copilot so it works inside Outlook Feeds decision02 Who gets which tools by default
  3. 3Team and unit requestsWhere personal use is turning into a rollout For exampleA department asking for the same tool for 40 people Feeds decision02 Who gets which tools by default
  4. 4Requests stopped by a ruleWhether the rule is clear, and whether a safe route exists For exampleStaff trying to summarize documents that contain SSNs Feeds decision03 What AI is never used for · 04 Sensitive requests
  5. 5Proposals by review levelWhether reviewers can keep up For exampleUnit chatbots waiting longer than the target time Feeds decision05 How new ideas get reviewed
  6. 6“Not the right tool” feedbackWhere people need training or a better match For exampleNew users matched to agent tools they can’t yet use well Feeds decision06 What training people need

Review it each term

  • Review the signals with the people who govern and support AI on your campus, such as IT, privacy, security, academic leadership and the units most affected.
  • Each signal ends in a decision: add, license, retire, explain a rule better, add reviewers, or add training.
  • Publish what changed, so people see their requests made a difference.
Owner: AI governance lead

Licensed CC BY 4.0: free to share and adapt with credit to Joe Sabado.

Find your AI tool

Tell us two things and we'll point you to the right tool, or the right person.

Step 1I am a…
Step 2I want to…

Pick the closest task. You can change it later.

Step 3Quick check: does any of this apply?

{{ start.err }}

Not sure? Treat it as yes. It only adds a few questions.

Starter kits

The fastest way to start. Each role group gets two tools the campus has already approved for everyday work, so you can skip the questions and start right away.

What's in a kit Two default tools with sign-in links, what each is good for, the kind of data you can use, and four simple guardrails.
Use it for Everyday work: drafting and editing, summarizing documents, meeting notes, study help, spreadsheets, brainstorming.
Don't use it when The work involves Restricted data, a decision about a person, or building something new. Get the right answer instead.

New to AI on campus? Read the “Before you start” checklist for your role.

Pick your role
All starter kits 1Path 1 · Starter kit

Starter kit: {{ kit.label }}

For everyday work, start with these. No form needed.

{{ t.name }} {{ t.desc }} Up to {{ t.clsLabel }}{{ i.x }} Who can use it: {{ t.ent }} Sign-in link: [placeholder]

Good for

  • {{ g.x }}

Data you can use

Up to {{ kit.clsLabel }}

{{ kit.flags }}

Guardrails

  • {{ g.x }}

Stop signs: don't use the kit if any of these apply

  1. Restricted data: health records, export-controlled research, card numbers, SSNs, legal material, confidential manuscripts
  2. A decision about a person
  3. Building, deploying or automating something
Intake questions
Question {{ intake.num }} of 4

{{ intake.title }}

{{ intake.help }}

No sign-in needed. Answers are counted anonymously to help the campus plan; nothing is tied to you. How requests are used

Do you have a Microsoft 365 Copilot license?
Staff, IT and leadership are assumed licensed. "Not sure" counts as no; the IT service desk can check.
Does the output decide anything about a person?
Does it include any of these? Check all that apply.
{{ intake.impliedNote }}
Do you have an approved research computing project?
Needed for the enclave and HIPAA use. "Not sure" counts as no; Research Computing can check.
Optional
How much have you used AI tools?
Agent tools need experience.

{{ intake.err }}

{{ intake.summary }}

Your answers stay in this page. Nothing is sent anywhere.

No result yet

Answer the five intake questions to see your match. If you followed a link, it may be incomplete.

 Your result
{{ result.res.pathNum }}Path {{ result.res.pathNum }} · {{ result.res.pathLabel }}
Recommended

{{ result.res.headline }}

{{ result.res.headline }}

 Alternative: {{ result.res.altLabel }} Last verified {{ result.res.verified }}
Also a good fit
  1. {{ m.rank }}{{ m.name }}Best for {{ m.best }}

Why this

{{ result.res.why }}

Try this prompt“{{ result.res.prompt }}”
Limitations{{ result.res.limits }}

Three tests

{{ t.name }}{{ t.sub }}{{ t.status }} {{ t.text }}

Conditions

  • {{ c.x }}
Review level: {{ result.res.bandLabel }}Decider: {{ result.res.bandDecider }}Target: {{ result.res.bandTime }}
{{ result.res.nextLabel }} Handled by: {{ result.res.who }}

Ruled out

Based on: {{ result.res.summary }}. Counted anonymously to help the campus plan; no names or content are kept. How requests are used

How this was decidedBefore you start
Was this the right tool?Anonymous and optional. {{ result.fbMsg }}
3Path 3 · Human consult

Get help

Some work needs a person before any AI tool. Find your situation below. Building a bot or AI feature? Propose it instead.

Not sure yet? Try before you choose. The campus AI platform lets you compare models side by side under campus data protections. Open the AI platform

{{ o.flag }}

{{ o.name }}

{{ o.situations }}

Bring
{{ o.bring }}
Outcomes
{{ o.outcomes }}
Contact
[placeholder]
Response
[placeholder]
4Path 4 · Propose something new

You have an initiative, not a tool question.

Building a bot or AI feature, deploying a model, or automating a decision about people goes through the campus’s proposal review instead of a tool pick. This demo uses the AI Strategic Compass, which scores each idea and sizes the review to the stakes, so small ideas move fast.

About the example: AI Strategic Compass

The Compass is a separate framework for deciding which AI initiatives a campus takes on. It scores each proposal, sizes the review to the stakes, and tracks results after launch. Picking a tool for your own work doesn’t need it; building something new does. A campus can plug in its own review process instead; the steps on this page stay the same.

Visit campusaicompass.com

Estimate your review level

Informational only. Reviewers score the five dimensions for real; one severe dimension can raise the band.
What data is involved?
Audience size
Does it judge or rank people?
{{ propose.band }}Decider: {{ propose.decider }}Target time: {{ propose.time }}
Answer all three to see an estimate.

What to prepare for Screen

  • A short description
  • The problem it solves
  • Domain: teaching, research, operations or engagement
  • Data involved
  • Who it affects
  • A rough draft of success measures
Start a proposalOpens the AI Strategic Compass, the example review process, in a new tab.

Tool catalog

Seventeen AI tools commonly used in higher education, shown as an example of how a campus might describe the tools it offers.

Listing a tool here is not an endorsement. The approvals, access and costs shown are illustrative, not any vendor’s or campus’s actual terms. The data levels (such as “Up to Confidential”) are for illustration only; check your campus’s data classification guidelines before using any tool. Product names belong to their owners.

{{ tools.count }}

No tool matches all of these filters. Clear one, or answer four questions for a recommendation.
ToolBest forSample promptWho can use itCostData levelProviderVerified
{{ t.short }} {{ t.best }} “{{ t.prompt }}” {{ t.ent }} {{ t.costCat }} Up to {{ t.clsLabel }} {{ t.provider }} {{ t.verifiedLabel }}

Which data each tool is approved for. Regulated data such as FERPA or HIPAA has extra rules; see each tool’s page.

Tool{{ cc.label }}
{{ t.short }} {{ m.mark }}
Tool details Open in new tab

{{ tool.name }}

{{ tool.desc }}

Listing a tool here is not an endorsement. The approvals, access and costs shown are illustrative, not any vendor’s or campus’s actual terms. The data levels (such as “Up to Confidential”) are for illustration only; check your campus’s data classification guidelines before using any tool. Product names belong to their owners.

{{ tool.why }}

{{ tool.provider }} · Last verified {{ tool.verifiedLabel }}
Open{{ tool.short }} Sign-in link: [placeholder]

Approved for

Up to {{ tool.clsLabel }}{{ tool.flags }}

Who can use it

{{ tool.ent }}License needed: {{ tool.license }} · Minimum skill: {{ tool.minProf }}

Works inside

{{ i.x }}Autonomy: {{ tool.autonomy }}

Strongest at

{{ tool.strong }}

Cost model

{{ tool.cost }}

Best for

{{ tool.best }}Try: “{{ tool.prompt }}”

Limitations

{{ tool.limits }}

The model, the tool around it, and the pair

The same model can be low-risk in one tool and high-risk in another. What the tool lets the model read and do matters as much as the model itself.

Model

Runs on
{{ tool.model }}
Model choice
{{ tool.choice }}
Version
{{ tool.pin }}
Model card
Vendor model card: [placeholder link]

Harness: the tool around it

Reads
{{ tool.reads }}
Can do
{{ tool.does }}
Who approves each step
{{ tool.approves }}
Memory
{{ tool.memory }}

This combination

{{ tool.combo }}

Up to {{ tool.clsLabel }}Autonomy: {{ tool.autonomy }}

Contract and controls

Contract terms
{{ tool.terms }}
Data handling
{{ tool.data }}
Admin controls
{{ tool.admin }}
Logs and records
{{ tool.logs }}
Accessibility
{{ tool.a11y }}
Quality for the task
{{ tool.quality }}
Who can get it
{{ tool.equity }}
Exit plan
{{ tool.exit }}
If something goes wrong
{{ tool.incident }}
Environmental footprint
{{ tool.green }}

Guardrails

  • {{ g.x }}

Example scenarios

  • {{ s.id }}{{ s.text }}

No examples yet.

Scenarios

500 sample campus requests, each run through the same rules as the matcher. Filter to find ones like yours, then open any one in the matcher to see the full result.

Scenarios are illustrative, not real requests. Tools, approvals and offices are examples, not any campus’s actual rules. Download all 500 (.xlsx)

{{ sx.count }}

Loading scenarios…

No scenario matches all of these filters. Clear one, or try your own request.
  1. {{ s.id }}{{ s.pathLabel }}{{ s.clsLabel }} {{ s.text }} {{ s.meta }}
    Recommended {{ s.toolName }} No AI tool for this Alternative: {{ s.alt }} · Handled by: {{ s.who }}
    Why, and the conditions
    {{ s.why }}Conditions: {{ s.conds }}Data: {{ s.flags }} · Course policy: {{ s.policy }}
    Try this in the matcher →

Before you start

A short checklist for your role, covering syllabus language, disclosure, data rules and accessibility. Open yours.

Faculty and instructorsBefore you use or recommend AI in a course
  • Put an AI statement in your syllabus. Say which uses are allowed and how students should disclose them.
  • Name the approved tools, and tell students that personal accounts are not allowed for coursework that includes student data.
  • Never put student records into a tool that isn’t approved for Confidential data.
  • Don’t use AI detectors as evidence of misconduct. Follow the academic integrity process.
  • Check the tool’s accessibility report before you require it.
  • Make sure every student can get the tool free. Don’t require a paid license.
Sample syllabus language
AI not allowedUsing AI tools on assignments in this course is not permitted unless I say otherwise for a specific task. If you’re unsure, ask before you start.
Limited usesYou may use campus-approved AI tools to brainstorm, get feedback or practice. Work you submit must be your own writing. Note any AI use at the end of the assignment.
Allowed with disclosureYou may use campus-approved AI tools in this course. Describe how you used them in a short note with each assignment, and check anything AI produced for accuracy.
Illustrative. Adapt it to your campus policy and your course.
StudentsBefore you use AI for coursework
  • Check your syllabus first. Rules differ by course and by assignment.
  • Use your campus account, never a personal one.
  • Disclose AI use the way your course asks.
  • Don’t upload other people’s personal information, or unpublished research without permission.
  • Check facts and citations yourself. AI tools can invent sources.
  • Need assistive AI for a disability? Disability Resources can arrange it with your instructor.
StaffBefore you use AI for your work
  • Use only tools approved for the data level you’re working with.
  • Don’t use AI to make or justify a decision about a person.
  • Redact Social Security numbers, card numbers and health details before anything goes into AI.
  • Keep the original in the system of record.
  • Talk with your supervisor and IT before automating a process.
  • Label AI-generated content you share publicly.

Ready? Find your AI tool or open your starter kit.

References

Public pages from colleges and universities that informed this design. Each one helps people match their needs to campus AI tools in a different way.

  1. Case Western Reserve University “Find your AI Tool”: three quick questions by role and task, with a note that answers aren’t collected. Used here for: Quick matcher, non-collection statement https://case.edu/ai/tools-resources
  2. USC AI Knowledge Hub A “What are you trying to do?” section and a detailed comparison table with sample prompts, best-for, cost, access and last-verified dates. Used here for: Sample prompts, last verified, limitations https://sites.usc.edu/ai-knowledge-hub/ai-tools/
  3. TU/e (Eindhoven University of Technology) Task-first navigation, role-specific guidance and privacy status badges, with recommended tools per task. Used here for: Role checklists, privacy badges https://tueindhoven.ai/education/aitools/index.html
  4. UCLA Digital & Technology Solutions An AI tools catalog and a campus comparison matrix. Used here for: Comparison table, data matrix https://dts.ucla.edu/products-services/ai-tools/campus-ai-tools-comparison
  5. UT Austin Enterprise Technology “Not sure which tool is right for you?”: comparison by capability, data use, licensing and use case. Used here for: Comparison by data use and licensing https://tech.utexas.edu/services-tools/ai
  6. Stanford University IT A gateway organized by provider, an AI Playground, and AI Catalyst consultations. Used here for: Playground, provider filter, consultation path https://uit.stanford.edu/ai
  7. University of Minnesota AI Hub “Which AI tools should you use?” with a role-tagged directory of tools reviewed for data handling. Used here for: Role tags on every tool https://ai.umn.edu/

Data behind this site

500 sample scenarios (Excel)Every request with its role, task, data level, matched tool, alternative, guardrails and reasoning, plus the tool catalog and matching rules. Illustrative.Download .xlsx Or explore them with filters →

Links checked October 2026. Listing a page here doesn’t mean that institution endorses this concept.

FAQ

Can I use my personal ChatGPT or other consumer account?

No. Consumer accounts can train on what you type and sit outside campus data agreements. Use your campus account; if no approved tool fits, the result tells you who to ask.

What counts as Confidential vs. Restricted data?

Confidential covers regulated or sensitive personal data: student records, personnel files, student financial data, identifiable research participants, donor data. Restricted is the highest-risk data: health records, export-controlled data, card numbers, SSNs, privileged legal material, confidential peer-review manuscripts.

Why was a tool I like ruled out?

It failed a pass/fail gate: its data approvals, your role's access, or the experience it needs. Your result lists the reason.

I need a license I don't have.

Ask the IT service desk. A license is worth requesting if the task is recurring.

Can AI help me grade?

Drafting feedback, yes, with conditions. Deciding grades, no: a human makes and documents the decision.

Does this site keep what I type?

No. Everything runs in your browser. A shared result link holds your answers in the link itself, not on a server.

Who decides what's in the catalog?

The institution, through its usual technology review: security, privacy, procurement and IT. Approved tools are recorded in the Campus AI Registry. The AI Strategic Compass is for new initiatives, such as building a bot or an AI feature, not for choosing a tool for everyday work.

I'm a novice. Where do I learn?

Campus AI basics training [placeholder].

About

Campus AI Match is a concept. It shows how a campus could turn its AI decisions into a simple guide for students, faculty and staff. It isn’t a product, and it isn’t connected to any campus’s real tools or policy.

The tool catalog shows AI tools commonly used in higher education as representative examples. Including a tool is not an endorsement, and the approvals, licenses and costs shown are illustrative, not actual institutional policy or vendor terms. Product names are trademarks of their owners.

This is a personal project. It is not associated with, supported by, or endorsed by my employer, and any views expressed here are my own.

Accessibility

This site aims to meet WCAG 2.2 AA. Everything works with a keyboard, color is never the only signal, and pages work at 320px wide and 200% zoom. If something gets in your way, contact [placeholder].

Concept by Joe Sabado. Licensed CC BY 4.0.
How it works

Four paths, and how requests travel them

Requests start with the simplest check and move on only when they need more. The first “yes” decides the path. If every answer is “no”, it’s something new that needs a proposal.

Why the tool matters, not just the AI model. Two tools can run the same model and still need different rules. A chat window only answers; an agent can open files, run code or send things on its own. What matters is what the tool lets the AI read, touch and do.

500 sample scenarios (Excel)Every request with its role, task, data level, matched tool, alternative, guardrails and reasoning, plus the tool catalog and matching rules. Illustrative.Download .xlsx Or explore them with filters →
  1. {{ t.n }}{{ t.label }} {{ t.gets }} Who decides: {{ t.decider }}
See the flowchart on the home page

Follow a request, start to finish

Four representative requests, one for each path. Pick one to see every step.

{{ walk.cur.who }} “{{ walk.cur.ask }}”
  1. {{ s.i }} {{ s.where }} {{ s.text }}
Who’s involved: {{ walk.cur.people }} Typical time: {{ walk.cur.time }} {{ walk.cur.cta }}

More examples

{{ e.who }} {{ e.ask }}
  1. {{ s.q }}{{ s.a }}
  2. {{ e.n }}{{ e.pathLabel }}
Answer: {{ e.result }} {{ e.why }}

Page not found

That page doesn't exist. Go home or find your AI tool.