Will adding generative AI to your data project improve the outcomes? It might, but it’s not guaranteed. It can be harmful and costly to add generative AI in the wrong contexts. But there is also great potential to create opportunities and efficiencies that may not be achieved without AI. To help you determine if your data project would benefit from adding a generative AI component, I have listed 9 questions below for you to consider.
- Do industry laws and organizational rules allow for the use of AI in the context of your project?
- Do you have, or can you hire someone with, the expertise to help you choose the correct AI model?
- Do you have high-quality data in a format that is consumable by your selected AI application?
- Can you afford the technology and administrative costs associated with the addition of AI?
- Do the end users of your solution know you are using AI? Do they have an alternate path to get what they need if they prefer to avoid AI or if the AI fails to deliver what they need?
- What happens when the AI output is wrong? Are people harmed or just inconvenienced? What is the cost in terms of time/money/brand reputation?
- Do you have a way to test the accuracy of the AI model output?
- Do you have a way to validate that adding generative AI improved the process?
- How can end users give feedback on the output from the AI model?
If you have more suggested questions, feel free to add them in the comments.
#’s 6, 7, 8, and 9! #3! And the others as well — all good questions. Being able to assess the cost/benefit is important in all projects, and rarely done, in my experience. And needs to be assessed also in my suggested #10: will using AI boost the marketability of your product, i.e., attract investors interested in firms using the latest technology? Will that be worth it?