SERVICE DESIGN JAM

2023

lobby of a design studio filled with people

Executive summary

I took part in the 2023 Service Design Jam at IBM iX in Berlin, a hackathon-style event with dozens of jammers and a great team of coaches and organizers. Six months into my transition from academia, it was my first experience with a full design sprint (exciting)!

My role: Member of the design team, carried out the entire double diamond design process in collaboration with 5 team members with various design backgrounds and experience (Damla, Melanie C, Melanie S, Sebastian, Sophie) with the guidance of Adam (coach).

Methods: Affinity mapping, guerilla interviews, heat mapping, “How Might We” problem mapping, idea generation, journey mapping, prototyping, user journey prototyping, user testing, Wizard of Oz testing

Tools: Post-Its, whiteboards, role playing, craft supplies

Duration: 48 hours

Deliverables: Persona, service pitch, prototype, demonstration

graph of the double diamond design process with service design methods mapped onto the appropriate stages

Methods we used in the double diamond design process

Main challenges

Our design challenge was the word “build:re:build” and an ambiguous video prompt full of shifting images of cities and unsettling atmospheric noises.

We only had 48 hours to go through an entire sprint, from ideation to demonstrating a solution.

The process

Design challenge

We convened in the workroom to decide on a team name (“Spice Girls and the Gang”) and three keywords based on the design challenge prompt.

the words AI, social spaces, and changeable on a whiteboard

Keywords: AI, social space, changeable

And that was it for Day 1!

The next morning, we started to more concretely map the problem we wanted to solve: How might we help people struggling to integrate in a world driven by technological change and capitalism find more secure livelihoods?

Research

Time to get some answers! We developed a short list of questions about how people feel about their jobs, the future, and AI and new technology, and took to the streets in pairs to conduct a round of guerilla interviews. As an anthropologist, I’m always curious to test my assumptions against what participants actually have to say.

We found a lot of ambivalence, even hostility, regarding AI and the digital world. At least half of our respondents seemed to have no clear sense of the benefits and drawbacks of AI. Many felt that AI and digital technology were likely to play an increasing role in the world of work, but that there could be unknown and serious consequences (the Terminator was mentioned more than once).

Bronwyn conducting a guerilla interview with a passerby

Conducting guerilla interviews

Define

Over lunch, we used our interview data to refine our How Might We problem statement: How might we help low-tech, but highly experienced workers prove their skills and qualifications in order to maintain their independence and reduce anxiety in an increasingly digitized world?

We sketched out a persona around this problem to provide an empathic anchor for our prototypes.

Adam (coach) pointing to personas on a whiteboard

Adam pulls out common themes from a persona brainstorming session

a sketch of a persona for a low-tech but experienced worker

Our team's first persona

Ideate

We then brainstormed some “What if …”, “Can we …”, and “I feel we should …” solutions and iterated on our ideas as a group with addendums and dot voting.

Ideating solutions

We split the team in two, with each group developing a user journey and a prototype, so that we could collect a greater variety of user feedback.

Prototype

Our group designed a prototype for a career event at a community center where low-tech job seekers, through a mix of human and AI assistance, could be matched with a mentor who would help them develop their digital skills. We wanted to emphasize in-person interactions and community during this user journey, given that the user had low tech literacy and might feel uncomfortable with too many digital interactions.

Damla, Sebastian, and I build our prototype for a digital skills mentoring meetup

The other group prototyped an AI-based video chatbot that conversed with the job seeker and helped them identify next steps for their job search. This group was excited about exploring the full potential of AI, especially given the surge of interest since the launch of ChatGPT.

Sophie shows off her group's prototype for an AI agent

Then it was time for the Idea Bazaar. Coaches, organizers, and fellow jammers toured a makeshift expo featuring the prototypes of all the teams so that we could test and gather feedback on the designs.

Our testers mentioned that they were sometimes confused by the number of steps and different people involved in the user journey. On the other hand, they appreciated the mix of human and AI guidance. “People need both,” said one tester.

Leading testers through our prototype

After the Bazaar, our team reconvened and each group shared our user feedback with the other. There was lots to marinate on as we wrapped up Day 2.

Iterate

Day 3 kicked off with some yoga, followed by an inspiring presentation from Andrea Goebel (IBM iX) on her team’s recent project for Lufthansa. It was great to get a better sense of what service design looks like in a professional context.

Back at our worktable, we had a lively discussion on how to proceed based on the user feedback and were having a little trouble letting go of our darlings. I was wary of continuing in the direction of an AI chatbot, given that our user wasn’t so comfortable with new technology, and of pushing through a hyped technology where existing (human) processes could already satisfy users’ needs and wants. Other team members pointed out that AI could be faster and more accessible in helping job seekers than many existing government services.

To help us converge on the next prototype, Adam had us each ideate and present service pitches, which we then voted on. We came together on a service pitch for supermarket cashier Beate, age 47.

Problem: Beate is at a loss after she loses her job due to automation, and the Job Center is unable to help her on short notice
Solution: The Job Center officer points Beate to a newly installed terminal with the Digital Job Companion (avoiding loaded terms like “AI” or “chatbot” in the name)
User and service actions: Beate goes to the terminal, and the Companion introduces herself and her function. There is an “escape” button if Beate wants human assistance. Beate tells the Companion about her work experience, skills, and goals. Based on this information, the Companion identifies possible careers, open positions, and opportunities for skill-building and networking
Happy ending: Beate goes to a networking event suggested by the Companion and meets someone who recommends her for a job, resulting in an interview

Service pitch

After all this and having lunch, we only had an hour to build our prototype and practice our demonstration! Time to delegate. Damla and Sebastian built the Digital Job Companion interface while Mel S, Mel C, Sophie, and I worked on a script for our respective roles as Beate, a narrator, a Job Center officer, and the Companion.

Conclusion

Present

Each team presented their results to the jam in a 5-minute demonstration. Sophie, as the Job Center Beamtin, stole the show.

Demonstrating the Digital Job Companion

Reflect

I enjoyed letting go of some of the perfectionism that I’ve learned in academia in favour of good-enough solutions and learning through failure.

I enjoyed both voicing my opinions and defending my viewpoints (re: AI, for example), but I committed wholeheartedly to the vision once the team decided on a direction.

As an ethnographer, I had obviously been looking forward to the research components of the jam, but I also really enjoyed communicating and ideating in a team setting. I was pleased to learn that many of the skills and principles I’d been practicing in my weekly improv theatre classes served me well, such as “Yes, and …” and killing your darlings for the sake of convergence.