Overview

At IBM, I’ve been very fortunate to have the opportunity to facilitate or co-facilitate numerous design thinking workshops. A new skill I’ve been developing though is around Design Thinking for AI and Machine Learning.

I’ve been directly training under Jennifer Sukis, design principal for AI and ML. (At IBM, a design principal is a director level position)

As of August 2019, I’ve directly facilitated 3 AI design thinking workshops, co-facilitated 2, and participated in 2.

Regarding traditional design thinking, I’ve led 5 and participated in 6.

Workshops I’ve been part of: 18 in 1 1/2 years.

I’ll walk you through the framework at a high level, but I’m happy to go into more detail and specifics if you contact me!

Framework

AI Design Thinking follows the IBM Enterprise Design Thinking for the most part, but with a few very specific differences.

Pre-Planning

Never forget the importance of pre-planning for a workshop! You’ve got to make sure you have all the right people there, goals for the workshop, and that you have already gotten everything you need to achieve those goals. Otherwise, you’re probably wasting your time and money.

User Needs

The first thing we do is orient ourselves around the user needs and intents. Following Indi Young‘s philosophy of inclusive persona design, we think about who a person is beyond their job title. We think about the roles they bring, whether they’re a troubleshooter, boss, snack enthusiast as well as a system administrator.

Data

Then, we take an inventory of all the data we have access to in the public and private spheres. Documenting that will help us later when we’re talking about AI solutions.

As-Is Scenario

An As-Is scenario map takes into account the current workflow that a user goes through.

Example IBM Design Thinking As-Is Scenario Map:

Brainstorm individually, step 2

^ This is the intent behind what we do, but we make some key changes in AI Design Thinking.

  1. We map out what the system is doing as well as what the user is doing.
  2. We specifically call out pain points in a separate row
  3. We have a row for AI opportunities
  4. Finally, a row for the data we’ll need for those AI opportunities.

We don’t focus on specific solutions for AI opportunities here, but we flag them.

Big Ideas

After mapping out the current scenario, we then reflect on the opportunities and pain points uncovered by the As-Is map.

We then spend time coming up with ideas specifically around those pain points.

To-Be + Technical Expert

We then reign in our big ideas and vet with with a technical expert to come up with a To-Be journey that incorporates our AI ideas.

Next Steps

After we’ve come up with our to-be, we get it on the roadmap! That might require making a press release, a prototype, or at the very least a presentation to execs. We align on who owns what and then we distribute that accountability to make sure it will become implemented.