Creating "Machine Learning Ready" Experiences that are Built to Improve

In January of 2018 I started a certificate program in Artificial Intelligence from Columbia University. Although the course itself is technical in nature, as I progress through I've started making connections to apply in the world of strategy. I'll be sharing these thoughts and observations on the connections between the technical material and business strategy here on LinkedIn and on my blog.

"A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E." -Tom M. Mitchell.

As my course has progressed we have finally gotten to Machine Learning, which is one of the most prominent approaches within Artificial Intelligence. In number 2 of my series I want to use the compelling operational definition of Machine Learning that Tom Mitchell created to talk about Experiences and how strategists, marketers, executives and creatives can adjust their approach to make sure the initial creation of the experience is one that can be improved upon and is therefore "Machine Learning" ready---essentially let's reverse engineer the definition to make sure our Experience Strategists are thinking about how to improve experiences in the future before they've even thought about that initial experience.

To do that you should first get comfortable with the definition from Tom Mitchell above...so let's break it down:

"A computer program is said to learn from experience E with respect to some class of tasks T ..."

We as strategists could take this to mean that the experience (E) we create will have some sort of technology (computer program) attached to it that will read tasks (T) that ultimately deliver the total of the experience.

Example 1 if we create a website (E) it will have images, copy and links (T) that our analytics program and CMS (computer program) will use to learn about the overall site experience (E).

Example 2 outside of the "digital" space is perhaps an in-store experience that we create (E) will have positioning of products, interactive kiosks and traffic flows (T) that will be monitored by cameras and barcodes on products which collect and store data (computer program) that learn about (E) from.

"...and performance measure P if its performance at tasks in T, as measured by P, improves with experience E."

Let's take this to mean that (P) or performance is only assigned at the task (T) level and not the experience level (E) as with the above definition (E) becomes the creative spark that determines success.

Example 1 Our website will have (P) attached to each (T) clicks on content, likes on images and click throughs on links that will depend on (E). Change the overall (E) and theoretically you've changed the (P) on (T). So before you implement an (E) make sure you think about the (T) that we can attach with (P) to make sure we are understanding the (P) of (E). In simple terms--- think about the components of your experience because that is the measurable part of your experience that allows a program to apply machine learning to creating a better experience.

Example 2 our in-store experience will have (P) attached to each (T) emotion on expressions our cameras detect, the number of barcodes scanned as product volume goes up which will depend on the (E) we create. The same simple terms apply as Example 1.

By complicating the simple idea of thinking about the tasks that are attached to the experiences it actually allows creatives and strategists to think more about how the initial experience they create will get better once it's created.