1 minute read

We started with 30+ AI use cases. The best 5 made the cut. Here’s how a curated approach sets up better outcomes.

How do you turn a sprawling list of 30+ AI ideas into a focused, high-impact portfolio? I recently helped a client do just that, using a stepwise process to move concepts from thoughtful brainstorms to a clear set of priorities.

Curating AI Use Cases

Filters for curating AI use cases

Here’s the methodology we applied:

𝗙𝗶𝗿𝘀𝘁, we ran a simple litmus test: Is this a problem that uniquely requires AI? About a fifth of the list was immediately re-categorized as systems improvements—important, but not for the AI team. That left us with ~25 contenders.

𝗡𝗲𝘅𝘁, we used a scoring rubric to evaluate each one on two axes: business impact (financials, customer satisfaction) and feasibility (cost, complexity). Clear criteria enabled consistent conversations across business units about the ratio of return to investment. Projects with the highest impact-to-investment ratio rose to the top.

𝗧𝗵𝗲𝗻, we looked for patterns, grouping use cases into functional themes like knowledge retrieval or visual rendering. This wasn’t just an organizing exercise; it was strategic. Building a capability in one area creates leverage for the next.

𝗙𝗶𝗻𝗮𝗹𝗹𝘆, we balanced the portfolio. We deliberately selected initiatives from different business units to broaden exposure and build enterprise-wide muscle. We also mixed iterative “quick wins” with a few strategic bets, mirroring the layers in my Enterprise AI Pyramid framework.

𝗧𝗵𝗲 𝗿𝗲𝘀𝘂𝗹𝘁: we curated the list from over 30 down to 5 high-potential projects ready for MVP design.

𝗧𝗵𝗲 𝗹𝗲𝘀𝘀𝗼𝗻: A successful AI strategy isn’t measured by how many projects you start, but by how many impactful projects reach production. Culling down to a strategic few ensures the right initiatives get the focus they need to cross the finish line.

What other criteria have you used in selecting AI use cases?

You May Also Enjoy