If you’re looking for an HCI elective to explore design innovation with artificial intelligence (AI), we recommend 05-317: Design of AI Products and Services. We’ve mentioned the course before in our newsletter here, but let’s take a deeper dive into the course content. I really enjoyed taking the class myself I’m excited to introduce one of the core concepts — the innovation framework of matchmaking.

What is matchmaking?

<aside> 👉

Let me first clarify that the matchmaking design innovation process is not referring to matchmaking algorithms in game design.

</aside>

Matchmaking, as used in the Design of AI Products course is the process of matching technology capabilities with activity domains to identify meaningful new applications for existing technologies.

Diagram with the four design steps in the matchmaking innovation process (1. technology characteristics, 2. work activities, 3. example sites, 4. domain characterization)

General design steps in the matchmaking innovation process to match technology to activity domains (Bly & Churchill)

The AI design course adopts the matchmaking framework as originally proposed by Sarah Bly and Elizabeth Churchill in 1999. Despite the date of the article, the process is timeless with technology innovations because it’s less about the technology itself and more about the process of how you apply it to create meaningful solutions across domains.

László Moholy-Nagy, a Bauhaus professor and early advocate for technology integration in art, said that “Design is the organization of materials and processes in the most productive way, in a harmonious balance of all elements necessary for a certain function.” Following Moholy-Nagy’s definition of design, it follows that: By seeing technology capabilities as materials, you can design new ways to organize them.

“We realized that there was a need for designers to peer in on this innovation process and work with AI, but they didn’t need to know the technical aspects of AI… They needed to look at AI in a really different way as a set of capabilities, as something that could be molded, much like clay or pixels or paper. Hence came this concept of AI as a design material, and that’s been the basis of some of my and my colleagues’ research around AI innovation. It’s also the foundation for a course… called the Design of AI Products and Services.”

— Jodi Forlizzi (Herbert A. Simon Professor of HCI), in an interview with The Informed Life

How we perform matchmaking in the course

Domains (left) matched with learning mechanism (center) and technical capabilities (right)

Domains (left) matched with learning mechanism (center) and technical capabilities (right)

When we perform matchmaking in Design of AI Products, we branch out from a single, specific technical capability to explore many potential applications. Specifically, we question each of the following (in order):

  1. Potential domains / activities: What are some industry areas where this technical capability might be relevant (education, entertainment, healthcare, etc.)? Alternatively, are there activities the capability enables that might apply across domains?
  2. Customers within each domain: Who are you helping? Would this be a business-to-business application or business-to-consumer?