Defining AI Use Cases

The promise of AI is alluring—there are so many exciting, intriguing possibilities (and attractive advertising), who can resist wanting to dive in? But if AI is going to be your business partner, what kind of partner will it be?

This is the question many leaders are asking.

AI offers such game-changing potential, it’s tempting to rush into a proof of concept with only a vague definition of goals. After all, the objective is to move fast and either fail fast and pivot, or succeed fast and repeat. This is the ideal state once you reach maturity, but it’s not the place to begin your AI journey.

As I covered in my last blog, success with AI is less about grand ideas (although those are necessary starting points) and more about well-defined use cases. Your use case drives your preparedness down a very specific path. Simply put, your ‘why’ defines your ‘what’, and your ‘what’ determines the ‘how’.

Here are five steps to help you think through and define an AI use case:

1. Consider the possibilities

   Start with a blank whiteboard. Look at the big challenges and opportunities in your business and industry. Explore what AI and machine learning can do to help overcome those problems or enable those opportunities. What are others in your industry doing? Include your key stakeholders—IT, Finance, Legal, HR, Contracts, Security, Lines of Business. Encourage outside-the-box thinking and bold ideas; empower people to think big.

2. Evaluate risk

   Every new technology introduces potential risk to your organization. Understanding AI—and the power it possesses—what risk do your potential use cases pose to your organization? What effect will they have on compliance, governance, HR, and legal requirements? How will they change operations? As a major airline recently learned, your organization will be held responsible for what AI does on your behalf—just like any other employee. Sorting through these questions will help you assess risk and feasibility of your proposed AI use cases.

3. Define ethical use

   What do you consider ethical use of AI? Where are your boundaries? How does that compare to industry and societal norms? What legal or regulatory requirements exist that will govern your use of the technology? What guidelines (beyond legal and regulatory) will you use to form a policy around AI and its use in your organization? These questions may drive prioritization and de-prioritization of the use cases you’re considering.

4. Review talent and cultural readiness

   AI presents a new way of working and interacting with colleagues and customers. It changes how work gets done and by whom and may be viewed as a threat for managers and employees who are still trying to understand the technology and determine how and when to use it. Embracing the skillset and cultural change is key to any successful AI use case.

5. Determine what is ‘in scope’ and ‘out of scope’
   Every AI use case needs a clear definition of scope. This starts with data. What datasets will be required to make your use case possible? Where is the data located? How is it governed? AI will put a magnifying glass on any data that has not been governed—it could expose datasets you might not want revealed. As you define the scope, be sure to include security in every step. Data security, privacy, compliance, and access control requirements all need to be considered as you formulate your use case.

Defining a use case (or multiple use cases) that aligns with your AI strategy is the first step to moving forward with AI. But as the saying goes, “you can’t eat the whole elephant at once.” If your use case is too broad or the scope is too big, it will be difficult to succeed in POC.

Narrowing your focus, being as specific as possible about what’s in scope and out of scope, and identifying clear outcomes will enable you to move to the next step in your AI journey: Assessing AI preparedness, which I will cover in the next blog in our AI series.  

For additional help, ePlus offers several advisory services—from envisioning workshops to infrastructure and storage assessments and data governance—specifically focused on helping organizations unlock the full capability of AI. Check out ePlus AI Ignite for more information.