AI Assessment Framework

I frequently engage with organizations about AI, focusing on how to achieve business objectives, measure ROI, and accelerate their AI journey, regardless of whether they are just starting out or already advanced.
One effective way to consider potential actions—and thus identify the associated blockers, risks, and benefits—is to plot scenarios using the AI Assessment Framework.

There isn’t a one-size-fits-all AI model. As I mentioned in earlier posts, there are thousands of models available, and it’s important to select the right one for your specific scenarios. For those on the Microsoft platform, you can take advantage of a comprehensive ecosystem of AI models without needing to switch clouds. This approach minimizes risks and costs while maximizing time to value. A visual of this ecosystem is here.

So to the Framework:

Think about use cases across 3 dimensions to determine Complexity, Risk and ROI. The three dimensions are:

Build or Buy
Internal or External
Automation or Manual

For example:

Build or Buy: How much Build (custom AI algorithms on Azure ML) Vs Buy (CoPilot for Sales) does your Use case need?
Internal or External: How Internally (Agent Assist) facing Vs Externally (Self-Serve) facing is it?
Automation or Manual: How automated (automated loan approval) Vs human manual Intervention (CoPilot for Sales)

The combination of these factors will lead to different approaches. For instance, a Buy + Internal + Manual approach involves a relatively straightforward implementation and distinct risk considerations, compared to a custom AI model that automatically approves loans.

Now as the visual indicates it’s not as binary as build or buy, each dimension has a scale, but this does help you think about how far along the scale for each dimension do you want or need to go.

Mark Margolis