— Practice · Advisory —

AI use case discovery,
that finds the real value.

AI use case discovery is how the Institute of AI finds the AI work worth doing and puts it in order. It maps where the value sits, scores every candidate on value and feasibility, and hands you a ranked shortlist, so you build the right thing first instead of the loudest thing first.

Book a discovery See the consultancy

— What it is —

Rank the opportunities before you build.

AI use case discovery is a structured process for finding the AI opportunities inside an organisation and ranking them by value and feasibility before anyone commits to building. It answers the question most organisations skip: of everything AI could do here, what is actually worth doing, and in what order.

The failure mode is rarely a shortage of ideas. It is starting with the idea that is loudest, newest or easiest to demonstrate, then discovering months later that the data was not there, the value was thin, or nobody owned the result. Building the wrong thing first is the expensive mistake. Discovery moves that judgement to the front, where it costs a conversation, instead of the end, where it costs a project.

The Institute of AI runs this as part of its independent practice: IoAI advises first and builds only what earns its place. Discovery is deliberately vendor-neutral. It ranks opportunities on their merits, never against a product IoAI happens to sell, so the shortlist you leave with is one you can defend to a board.

— The value by feasibility grid —

Four quadrants, one clear order of play.

Every candidate lands somewhere on two axes: how much value it would create, and how realistic it is to build well. Where it lands decides what IoAI advises doing with it.

HighValueLow

High value · Lower feasibility

Big bets

Genuinely valuable opportunities held back by missing data, unclear ownership, or integration and governance gaps.

IoAI advises

Sequence, do not abandon. Clear the blockers deliberately so the value becomes reachable, rather than forcing a build before the ground is ready.

High value · High feasibility

Quick wins

Well-scoped tasks with clear value where the data, tools and skills already exist to deliver them.

IoAI advises

Start here. Move a small number straight into a proof of concept for an early, visible result that earns confidence for everything after it.

Lower value · Lower feasibility

Money pits

Hard, costly ideas with thin returns, often the impressive demo that caught an eye in a meeting.

IoAI advises

Say no, in writing. The most valuable output of discovery is frequently the list of things that are not worth building.

Lower value · High feasibility

Fill-ins

Easy, low-risk automations that are cheap to stand up but move the needle only a little.

IoAI advises

Do them opportunistically, never ahead of the quick wins. They build momentum and everyday AI literacy rather than headline outcomes.

LowFeasibilityHigh

— How we score —

Two axes decide what makes the shortlist.

The grid is only as honest as the scoring behind it. Each candidate is rated on the same plain criteria, so the ranking holds up when someone asks why.

The vertical axis

Value

Size of the prize: Time saved, cost avoided, revenue enabled or risk reduced, expressed in terms the business already tracks.
Frequency and reach: How often the task runs, and how many people or customers each run of it touches.
Strategic fit: Whether it moves something the organisation genuinely cares about, not just something that is easy to count.
Quality of outcome: Fewer errors, faster answers and a better experience for the people on the receiving end of the work.

The horizontal axis

Feasibility

Data readiness: Whether the information the use case needs exists, is accessible and is good enough to rely on.
Technical fit: How well the task suits what AI does reliably today, and how cleanly it joins your existing systems.
Skills and ownership: Whether someone can run it, maintain it and be accountable for it once the initial attention moves on.
Risk and governance: Data protection, the tolerance for error, and the controls the use case would need to be defensible.

— How the discovery runs —

From the whole map to a costed shortlist.

A short, focused engagement that reaches decisions quickly. Four steps take you from everything AI might do to the handful worth starting.

Map the work

Break the organisation into its real processes and tasks, so candidates come from where work actually happens, not a wish list.

Gather candidates

Surface every plausible use case from teams, live pain points and places AI is already proven, without judging them yet.

Score

Rate each candidate on value and feasibility against clear, agreed criteria, so the ranking is defensible rather than opinion.

Shortlist

Plot the scores on the value-by-feasibility grid and agree a sequenced shortlist: build first, prepare, or drop.

— What you walk away with —

A shortlist you can defend.

Discovery ends in decisions, not a shelf-bound report. Everything below is written down, ranked and ready to act on.

A ranked use case register: every candidate scored on value and feasibility, and plotted on the grid.

A sequenced shortlist: the few worth doing first, with the reasoning written down so it survives a change of mind.

An explicit not-now and not-ever list: the ideas discovery has ruled out, and exactly why.

The blockers to clear: the data, skills and governance gaps standing between a big bet and a real build.

A costed next step: usually a scoped proof of concept on the strongest quick win.

— Common questions —

AI use case discovery, the common questions.

What is AI use case discovery?+

It is a structured process for finding the AI opportunities inside an organisation and ranking them by value and feasibility before any building begins. Rather than starting from a tool and looking for somewhere to use it, discovery starts from the work, surfaces every plausible use case, scores them on clear criteria, and hands you a sequenced shortlist of what is genuinely worth doing.

Why not just start building the obvious idea?+

Because the most expensive AI mistake is building the wrong thing first. The obvious idea is often the loudest or the easiest to demonstrate, not the one with the best return or the readiest data. Discovery moves that judgement to the front, where changing your mind costs a conversation, instead of the end, where it costs a project.

How long does a discovery take?+

It is scoped to the size and complexity of the organisation and kept deliberately short and focused, because its job is to reach clear decisions quickly, not to produce a lengthy report. The Institute of AI agrees the scope up front, so you know exactly what the engagement covers before it starts.

What happens after the discovery?+

You leave with a ranked register and a sequenced shortlist. The usual next step is to take the strongest quick win into a scoped proof of concept, so the value is tested in practice before it is scaled. IoAI can advise only, or advise and then build what it recommends.

Do we need an AI strategy first?+

No. Discovery and strategy reinforce each other: a grounded view of which use cases are worth pursuing is often the most useful input into an AI strategy, and an existing strategy gives discovery its priorities. Either can come first, and IoAI can help with both.

— Practice —

Find the use cases
worth building.

Talk to the Institute of AI about a structured AI use case discovery: a value-by-feasibility ranking of your opportunities and a shortlist you can take to a board, before anyone writes a line of code.

Book a discovery See the consultancy

— Go deeper

— Advisory