Apply · The planning process

AI strategic planning

Most campuses don't lack AI ideas: they lack a plan that leadership can endorse, a strategist can see a path through, and a practitioner can act on Monday. This is the method for producing one: a repeatable planning cycle that turns the framework's eight pillars and tools into a single arc, ending in a written strategy with a budget, an owner, and a date. It's a guide to running the process, not a template to fill in.

The starting condition

On most campuses, AI arrives long before any plan does. Left alone, the pattern is consistent:

Everyone owns it, so no one doesAI work is underway in courses, research, IT pilots, and student services, but no one is accountable for the whole.
Effort duplicates quietlyUnits solve the same problems in parallel without knowing it.
Shadow and vendor use spreadsTools switch on by default and staff adopt AI informally, faster than policy gets written.
The public face drifts out of syncWhat the institution says about AI stops matching what it's doing.

Planning turns that scattered activity into one owned direction. The rest of this page is how.

What you're producing

Why this mattersA good AI plan is six things on one page, not a hundred. Define the target first, and every phase below has something concrete to build toward.
01

A purpose tied to mission

Why AI, in your own words: the outcomes it serves for students, research, and operations. A mission statement, not a technology one.

02

Commitments you'll be held to

The principles that bound every AI decision (equity, privacy, transparency, human judgment), stated so a board, senate, and the public can hold you to them.

03

An accurate baseline

Where you stand across the eight pillars, and what AI is already doing on campus, including shadow and vendor use no one approved.

04

A prioritized portfolio

The few initiatives you'll pursue now, in order, and what you're deferring and why.

05

Governance & decision rights

Who decides what, who is accountable, and how a new AI use gets reviewed. The plan names owners, not just ambitions.

06

Resources & measures

What it costs, who owns each piece, and the leading indicators that show it's working before long-term outcomes land.

The planning arc

Six phases, each producing one concrete output and handing off to the framework tool that does the work. Adapt the order to your context: a resourced campus runs several in parallel; a nascent one may spend most of its first cycle in Phases 1–3.

1
Phase 1

Convene & set the mandate

A plan without a sponsor and a team is a document nobody owns. Name who is accountable and what's in scope before discussing a single initiative.

In the roomA cabinet-level sponsor and a small, cross-functional planning team
What you do
This producesA chartered planning team, a named sponsor, and an agreed scope.
Decide now that AI runs through the whole plan (teaching, research, operations, public engagement), not a standalone technology chapter.
2
Phase 2

Establish an accurate baseline

Plan from where you are, not where you wish you were. Get a truthful read of your maturity and of the AI already running on campus.

In the roomThe planning team with IT, data, and unit representatives
What you do
This producesA baseline maturity read and an inventory of current AI use.
3
Phase 3

Set direction

Decide what you stand for before choosing initiatives. Direction keeps a hundred small AI decisions coherent.

In the roomThe sponsor and planning team, with a faculty and shared-governance voice
What you do
This producesA short purpose statement and a set of principles on the record.
Name your posture in plain verbs (explore, accelerate, safeguard). Shared, action-oriented words make later decisions easier to coordinate and explain.
4
Phase 4

Prioritize across the pillars

Decide which few pillars matter most now, and score candidate initiatives so choices are defensible, not loudest-voice-wins.

In the roomThe planning team with unit sponsors who own the initiatives
What you do
This producesA ranked shortlist of initiatives, with a clear rationale for what you defer.
5
Phase 5

Draft & pressure-test

Write the plan around the six components above, then test it against real risks, recognized standards, and the people who will live with it.

In the roomThe planning team with legal, IT security, and stakeholder representatives
What you do
This producesA drafted plan that has survived a first round of scrutiny.
6
Phase 6

Resource, approve & sequence

A plan becomes real with a budget, an owner, and a place on the calendar. Endorse it, fund it, and schedule the first moves so work starts the week it's approved.

In the roomThe sponsor with cabinet or board, and finance
What you do
This producesAn approved, resourced plan with the first 90 days already scheduled.

Every institution starts somewhere different

There is no single on-ramp. The same arc flexes to your institution: its type, its maturity stage, and the proportionality of a response that actually fits your risk and capacity. Find the nearest example and read across all three; the point isn't to match it exactly, it's to right-size your own plan.

Institution typeCommunity / access-focused college
Maturity stageInitial · Ad hoc
ProportionalityFeatherweight: one coordinator, borrowed templates, no new office
You're likely here if
  • AI is already in classrooms and back offices, but no one owns the whole picture.
  • Nothing is written down; what happens depends on individuals.

Weight the front of the arc. The win is a named owner, one piece of published guidance, and a stated direction, not a full plan. Don't build an apparatus you can't staff.

Where the arc leansPhases 1–3 · Convene → Baseline → Set direction
Your first moveTake the quick read, then name one AI “front door”Find your starting point
Watch for: Governance heavier than you can sustain. One page of guidance beats an unstaffed committee.

From plan to practice

Approval is the start, not the finish. The plan becomes an operating rhythm: monthly intake, quarterly portfolio reviews, an annual reset, and you return to this cycle each year with a fresh baseline as the program matures. Plan once; run it continuously.

Where AI plans go wrong

Each of these has stalled a good plan on a real campus. The framework doesn't remove them: it names them so you can plan around them deliberately.

Planning to perfection
“We'll publish once it's comprehensive.”

Publish a v1. A plan people can act on this month beats a complete one no one has seen.

A plan with no owner
“Everyone agreed it was important.”

Put a name next to every initiative. Enthusiasm without an owner is how plans stall.

Borrowing another campus's priorities
“This worked at a peer institution.”

Borrow their structure, not their priorities. Plan from your own baseline, mission, and risk.

Procurement before direction
“Let's buy the platform, then write the plan around it.”

Direction decides the tools. Set purpose and priorities first, then let them constrain what you buy.

Skipping the baseline
“We roughly know where we stand.”

Run the assessment anyway. Without it you plan for the campus you picture, not the one you have.

A public face before a plan
“Let's announce the AI vision, then work out the details.”

Set direction and ownership first. A public AI profile with no plan behind it is out of date the day it's posted.

A plan is a decision, not a document

Run the arc once and you have a strategy your cabinet endorsed, your strategists can sequence, and your practitioners can start on. Then it becomes a rhythm.