Identify a use case for AI

1. Identify a use case for AI

Welcome back.

2. Five principles of a good use case

After talking about benefits, limitations, and use cases for an AI solution, let's talk about five principles for picking a great use case to start with. First, start with the "Why".

3. Five principles of a good use case

Augment employees, don't build AI in replace of them.

4. Five principles of a good use case

Always make sure to align with the business goals and vision.

5. Five principles of a good use case

Make sure sufficient data is available.

6. Five principles of a good use case

Finally, deconstruct into small tasks that can be automated or augmented.

7. Why build an AI solution?

It's exciting to dive into specifics about what, where, and how. But, it's important to first figure out "Why".

8. Why build an AI solution?

Why should AI be used for this problem?

9. Why build an AI solution?

Why will it improve the end-user experience - the employee or customer?

10. Why build an AI solution?

Why will it be important for the business?

11. Why build an AI solution?

Why now? Understanding why a problem needs to be solved and why AI is the way to solve it will make a use case stronger as well as direct the other principles.

12. Augment not replace

Secondly, AI is designed to work alongside humans - not replace them. It should assist and enhance our abilities.

13. Alignment with business goals

Next, find a use case where impact can be tied to a business metric. This will ensure the initiative is supporting the organization towards its vision. Similarly, evaluate the potential cost and risk to the business at each stage - proof of concept, full implementation, and governance.

14. Data, data, data

Moving to data - we've mentioned this a couple of times, but make sure there is enough data for the AI solution to learn from.

15. Deconstruct into small tasks

Finally, most roles have some aspect in which AI could be applied. After identifying one or more possible use cases, deconstruct the job, or process

16. Deconstruct into small tasks

into smaller tasks. This can help assess which tasks are suitable for an AI solution. Tasks that are well-defined or involve predicting an outcome, scoring, summarizing, labeling, or content-creation - are good places to start.

17. Don't use AI for this

With these principles in mind, there are some initiatives that can be solved without the use of AI. First, don't build a complicated solution to accomplish something minor or simple. ChatGPT isn't necessary for simple spell checking,

18. Don't use AI for this

and a deep learning model shouldn't be used for simple, one-time calculations.

19. Don't use AI for this

Instead, use generative AI to automate regular calculations of metrics and forecasts,

20. Don't use AI for this

write a small report,

21. Don't use AI for this

and create visualizations showing these trends. This is an example of using AI to rethink the process for creating and sharing insights about the business. Areas involving ethics and judgments, personal information or high-emotion situations, should be avoided.

22. Questions to identify a use case

Here are some other questions to ask to identify use cases. How are competitors using AI? What about partners in adjacent, or unrelated, industries? What are some areas of the business that are in clear need of support? Where would there be funding to support a small scale project, or a full scale implementation?

23. Start small with a POC

Once use cases are identified, the next step is to build a proof of concept. Keep in mind the power of a small scope at first. A small-scale POC

24. Start small with a POC

with small-scale impact

25. Start small with a POC

can facilitate big learnings. This is a great way to test, improve, and repeat until a solution is strong enough to scale.

26. Small-scale impact but big ROI

The POC may have small-scale impact, but it should tie to a potential for high return on investment, or ROI. Think of the POC impact as a sample of what the full solution could be. Return can be measured in revenue, time, retained customers, or any other important business metric.

27. Let's practice!

Before moving on to building a POC, we will identify suitable use cases!

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.