Insights from Irv Lustig: Risk Assessment for Sales and Fast Prototypes
Irv Lustig covers a lot of ground in our interview about the Princeton Twenty. The Princeton Twenty is a framework for understanding success and risk factors in the deployment of optimization and AI projects.
Here are two lessons I took from the interview:
One, a risk assessment tool for an optimization or AI project is about more than just risk assessment.
Irv talked about how the tool can be used in the sales cycle to give the client confidence that you can deliver the project, to help you properly scope the amount of work, and even to decide if you should walk away from a project (if you think there is more risk than the client does.) Irv discussed this as an external consultant, but the same lessons would apply internally.
Additionally, the tool can be used after the project to structure the review and help you improve on future projects.
Two, it is important to build a rough interface to show the client the types of decisions the system will make.
One way to reduce risk is to show the client, early on, the types of decisions that are made. You can’t wait for a production-ready interface to do this. You need to build an interface and present the solution to clients quickly.
When you do this, you get great feedback: you identify errors in the data, clarify objectives, and gain user buy-in.
I liked this point because I stress it in my capstone class: show your initial results early.
There are many other great insights. You can find the interview on YouTube, Spotify, or Apple Podcasts.
If you want to join the podcast’s mailing list, send an email to decisionintelligencepodcast@gmail.com.
You can find the episode on YouTube, Spotify, or Apple Podcasts.