Where should a university actually use AI?

Universities do not have a shortage of AI use cases.

Admissions, teaching, assessment, student support, career services, research, marketing, alumni engagement, administration. The list can become very long very quickly.

That is precisely the problem.

The important question for institutional leaders is not:

Where can we use AI?

It is:

Where should we use AI?

AI is a means to an end. The end is an institutional outcome that matters.

That distinction should shape how universities think about AI.

Start with the outcome

A common starting point is to ask what AI can do.

What can be automated? Where can an AI assistant be introduced? Which decisions can be supported by prediction? Which processes can be personalised?

These are useful questions, but they come too early.

The better starting point is:

Which institutional outcomes do we most need to improve?

For most universities, the major opportunities will fall broadly into three areas: student outcomes, institutional productivity and growth.

Once the outcome is clear, the role of AI becomes much easier to evaluate.

Take admissions. "Using AI in admissions" is not an outcome. The institution may actually be trying to improve the quality of the incoming cohort, increase conversion among desirable applicants, reduce acquisition cost or make better admission decisions.

AI is useful only if it can materially improve one or more of those outcomes.

The same applies to career services. The objective is not to become "AI-enabled." It may be to improve student-role matching, identify capability gaps earlier or strengthen employment outcomes.

The shift is simple but important:

Do not begin with "What can we do with AI?" Begin with "What result are we trying to improve?"

Where AI can create value

Student outcomes

AI can potentially improve decisions and interventions across the student journey, from recruitment and admissions to learning, progression and employment.

The strongest opportunities are likely to be those where better prediction, personalisation, information or decision support can materially improve a student outcome.

The objective is not to put AI into every student interaction. It is to identify the points where it can make a meaningful difference.

Institutional productivity

AI can also increase the productive capacity of the institution.

This is broader than automating routine tasks.

Universities contain fragmented information, repetitive work, manual hand-offs and institutional knowledge that is often difficult to access. AI may help reduce this friction, simplify processes and allow faculty and staff to spend more time on work where human judgment adds greater value.

Automating an inefficient process, however, does not necessarily create much value.

The relevant question is not simply what AI can automate, but what additional capability or effectiveness the institution gains.

Growth

AI may also strengthen the institution's ability to grow.

That could involve student recruitment, programme development, executive education, corporate engagement, alumni relationships or new delivery models.

Again, activity should not be confused with impact.

More leads, more content or more outreach matter only if they translate into outcomes such as better student quality, stronger conversion, viable programmes, higher contribution or deeper institutional relationships.

AI should support the growth strategy, not substitute for one.

Choosing where AI matters

Starting with outcomes does not mean that every AI opportunity linked to an important outcome deserves investment.

Institutions still need to judge whether the potential impact is material, whether the necessary data and organisational capabilities exist, whether the risks are acceptable, and whether the result can be measured.

That requires selectivity.

A few well-chosen initiatives that materially improve student outcomes, institutional productivity or growth may create far more value than a large portfolio of disconnected AI pilots.

The objective is not to maximise the use of AI. It is to identify where AI can make enough of a difference to deserve institutional attention.

Begin with the institution

The most useful starting point for an AI strategy may therefore be a set of questions that barely mention AI.

Where are student outcomes falling short? Where is institutional effort producing too little value? Which important decisions are being made with inadequate information? Where is growth constrained? Which priorities could be materially improved through better prediction, personalisation, knowledge access or decision support?

Only then should the institution ask what role AI can play.

AI is creating remarkable new capabilities, and higher education should take them seriously. But universities will create the most value when they resist the temptation to use AI simply because they can.

Start with the outcome that matters. Then ask whether AI can materially improve it.