No organisation can afford to sit back and ignore the potential of AI. Yet, rushing to adopt AI just because the technology exists can be just as damaging to a business. To fully realise the potential of AI – indeed, any new technology – you need to approach it with a clear business goal or need in mind. In other words, you need to identify and narrow down to the most valuable AI opportunities for your business (which means your approach could be quite different from that of your competitors).
When I work with organisations to develop their AI strategy, we break the process down into two fundamental steps:
- First, you need to identify the potential applications (use cases) of AI in your business.
- Then you can begin to whittle those potential use cases down to just a few top priorities.
Let’s explore each step in a little more detail.
Step 1. Identifying your AI use cases
Linking your AI strategy to your business strategy is the best way to ensure AI delivers maximum value for the business. Therefore, this first step involves looking at what your business is trying to achieve, and what unique challenges your business is facing, and then identifying potential solutions through AI. What you’ll end up with is a list of potential AI projects or use cases. An AI consultant can help you with this process.
Don’t limit yourself to a small number of use cases at this stage, as you’ll be narrowing down your options in the second step. For now, your goal is to explore the many ways in which AI could help your organisation achieve its key strategic goals. This might include:
- Making your products more intelligent
- Developing more intelligent services
- Building a deeper understanding of customers
- Making business processes more intelligent
- Automating core business functions
- Automating time-consuming, repetitive or mundane tasks
Having pinpointed some potential uses in your business, you can now begin to flesh them out in more detail. Consider the following factors for each potential use case (this ensures you’re approaching each use in the same structured, thorough way):
1. Link to strategic goal. Because it’s so easy to get caught up in all the incredible things AI can do, it’s vital you can link each use case to a strategic business goal. If a potential use of AI doesn’t link to your overarching business strategy, I’d question whether it’s really worth the expense and disruption. So ask yourself, “How would this use of AI help the business achieve its objectives, grow and prosper?”
2. AI objective Here, you can define your AI-related goal in more detail. What, specifically, are you hoping to achieve through this use of AI? For example, you may be looking to solve a particular business-critical problem, reduce employee turnover, improve health and safety, etc.
3. Measures of success (KPIs). This section answers the question, “What does success look like for this AI project, and how will we measure success?” Therefore, identify which business metrics/KPIs you could use to track progress against your AI objective.
4. Use case owner. Who in the business would be responsible for this AI use case and assume overall ownership of the project?
5. AI approach and data required. Which AI approach (for example, machine learning, deep learning, computer vision) is required to achieve your AI objective, and what sort of data do you need? Again, you may need to enlist the help of a data/AI consultant throughout this process, if you don’t have AI capabilities in house.
6. Ethical and legal issues. Here, you’ll need to consider the potential legal implications of your AI use case (including consent and data privacy), as well as the ethical implications.
7. Technology and infrastructure. It’s very likely each use case will require some technology and infrastructure changes. Here, you consider what systems, software, and hardware you might need to achieve your AI objective.
8. Skills and capacity. Here, consider the skills gaps that might prevent you from achieving your AI objective – and how you will close those gaps. This may involve training staff, hiring new talent or partnering with external providers.
9. Implementation. This is where you set out the potential implementation challenges and roadblocks that you’ll need to overcome if your AI objective is to become reality.
10. Change management. This is normally lumped in with implementation, but I tend to separate it out to emphasise how vital it is to manage AI-related changes carefully – after all, AI may lead to big changes in how your business operates. For example, if you’re automating or streamlining processes, this may impact the work of your employees, so how will you manage this while promoting a positive AI culture?
I like to use a formal template to help capture all these details, but you can do this however you like. Having evaluated each use case in detail, you’re now ready to narrow down your options. Which brings me to…
Step 2. Working out your AI priorities
You may have identified as many as 10 or 15 use cases in the previous phase, but trying to embark on too many AI projects at once can spell disaster. That’s why you now need to rank your use cases in order of their strategic importance to the business.
Your goal here is to arrive at:
- Your top one, two or three key strategic use cases. These are your absolute top AI priorities for now – the AI use cases that represent the biggest opportunities for your business, or will help solve your biggest business challenges. If you’re a smaller business, you may only want to focus on one key AI priority at a time. For a bigger business, two or three should be doable. But don’t be tempted to prioritise more than three.
- One or two AI “quick wins”. From your use cases, it often helps to identify one or two short-term, smaller AI projects that are relatively quick, easy, and inexpensive to implement. Identifying these quick wins will help you demonstrate the value of AI, win people over, and sow the seeds for your bigger AI projects.
As for those use cases that don’t make the cut for now, don’t discard them completely. As you achieve one AI project, you may find you have capacity to move onto another use case – or it may be that your overall strategic priorities shift, which means AI use cases that were identified some time ago are no longer relevant. I, therefore, recommend you repeat this process of identifying and prioritising AI use cases at least once a year, or every time you review your overarching business strategy.