In this article: Identifying a practical first use case by focusing on a real problem, a clear outcome and work that genuinely matters. | 7 minute read.
It’s 2026, and you’re already using AI – drafting emails, summarising documents, creating learning content and trying the AI features appearing in the systems you use every day.
But what should you actually be using AI for?
The technology is developing so quickly, and there is an understandable temptation to start with the tool. A new product is demonstrated, a feature is released or somebody sees an impressive example and asks: “Could we use that here?”
Before long, the organisation can find itself searching for a problem that suits the technology. Sound familiar?
A more sensible place to start is with the work itself.
Where are staff losing time? Where are your learners having to wait? Where does important information get missed? Which activities – whether for learners, tutors, or back office staff – are repetitive, inconsistent or just harder than they need to be?
Once the opportunities to improve are clear, it becomes much easier to decide whether AI can genuinely help.
Begin with a problem people (your people!) recognise
Let’s imagine that, on average, tutors in your organisation spend 20 minutes preparing for every progress review.
They gather the information they need from several places. They check a learner’s recent activity, previous review notes, any evidence that’s still outstanding, recent employer feedback, and the learner’s upcoming milestones.
A clear opportunity for here is creating straightforward efficiency - tutors already know how to prepare well for a review. But they’re spending a disproportionate amount of time assembling the information they need before they can apply their own expertise in preparation.
So rather than thinking broadly: “We’d like our tutors to use AI more”, let’s start identifying and articulating the opportunity more clearly: “Our tutors spend a lot of time gathering information prior to holding progress reviews, leaving them less time to prepare meaningfully for these learner conversations.”
From there, we can start to consider: “Can AI help by bringing the relevant information together? By potentially producing a useful first summary as a faster starting point for a tutor?”
Now, you have a defined use case relating to a real task that affects a clear group of people in your organization and will produce an outcome you can definitively assess.
It also helps you to keep the ultimate purpose of the technology in perspective – the goal is of course not to replace the tutor’s preparation or professional judgement, but to reduce the work involved in locating and organizing the information they need.
Consider what, specifically, should improve
Broad goals such as “improve efficiency” or “enhance the learner experience” sound positive, but are notoriously difficult to evaluate.
For this progress review example, goals might be one, some, or all of the following:
- reduce the time tutors spend preparing
- improve the consistency of the information tutors consider when preparing reviews
- make overdue actions or unresolved concerns easier to identify
- prevent any (relevant) employer feedback being overlooked
- help tutors to focus a greater proportion of the review conversation on the learner’s next steps.
The same principle would apply if you’re considering AI technology to amplify learner support.
“Use AI to support learners” is very broad – and difficult to set-up and assess.
A clearer, more specific use case might be:
Learners who get stuck outside working hours have no immediate way to ask for more clarity on the learning activities they are trying to complete.
This use case would lead to a defined application of the AI technology: providing a first line of support that will helps learners understand what an activity require or find relevant supporting information during moments when their tutor is unavailable (enter: Bud Assist for Learners of course!)
Remember, as we explored in our article on why AI assistants are only as useful as the information they can access, the true effectiveness of that support will depend highly on what the AI knows about the learner, their programme and the activity they are completing. But the starting point remains the same: a recognisable problem that matters to the people who are experiencing it.
Don’t overlook those everyday tasks
Your best first AI use case probably won’t feel transformative – and that’s OK!
In fact, moments of friction that happen repeatedly are more likely to create much more cumulative work than larger, less frequent challenges might.
Some possible starting points to consider might include:
- preparing a first draft of routine learner communication
- summarising relevant information before a review
- highlighting learner records that may need attention
- helping learners when they become stuck.
Something that saves one person ten minutes might not feel dramatic, but repeated across a large team, several times a week, can become really significant.
Starting with a really specific, focused use case will also give your organisation the opportunity to learn. You can see how staff use the capability, where AI output needs any improvement and whether it creates meaningful value in practice. Those learnings are much more useful than launching a significant AI programme without understanding how the technology will fit into everyday work.
Make sure the AI has what it needs
Once you have identified a possible use case, ensure the technology can access the information required to perform the task properly.
To create a useful progress review preparation summary, for example, the AI might need to understand:
- which activities the learner was expected to complete
- what they have actually completed since the previous review
- which activities or evidence remain outstanding
- what has been previously discussed and agreed
- whether the learner’s employer has provided any relevant feedback
- which milestones or deadlines are approaching.
Where any information is incomplete, inconsistent or spread across disconnected systems, the usefulness of the output will inevitably be limited.
That’s why AI adoption can’t be separated from your technology and data foundations – a connected environment gives it a much better chance of understanding the wider context of the work it is being asked to support.
In summary: Choose the work before you choose the technology
It’s 2026: There will always be a new tool, a new feature or a more ambitious demonstration of what AI might be able to do.
But remember that the strongest starting point is often less exciting on the surface – just a real piece of work that people recognise, a problem they would genuinely like to solve and an improvement you can describe clearly.
Start there, and then check whether AI is the right way to help, whether it has the information it needs and whether the result will make the work noticeably better.
Choosing the right first use case is not about finding the most impressive application of AI. It is about finding a useful place to begin.
Keep an eye out for our next article in this AI Starting Point series: Implementing and assessing the success of your first AI use case.
.svg)
-3.png)