Participant Activity Book

AI-Enabled Service Design Thinking

Four clusters of team activities that take one RTA service from a messy current journey to a tested future concept. Every step gives you what to do, what it needs from earlier steps, a copy-ready Claude prompt and the table your output should fill.

Start with Step 1Set up your team
Two journeys: the top one full of confusion and error markers, the bottom one smooth with green checkmarks ending in a satisfied user.
Current stateFuture state

Master instructions

  1. Work as a team, with one AI thread. Keep every activity in a single Claude conversation so context carries forward.
  2. Feed forward. Each activity needs the dataset or outcomes from earlier steps. Check the Needs line on every step.
  3. Stick to the output format. Share the table with Claude so results are easy to review.
  4. Treat AI output as a first draft. Review, reflect and agree as a team. Your judgement is the crucial part.

The prompts are designed to run in order, but you may apply your own judgement as long as the outcome matches the output format with all relevant details.

Claude labels each finding

EVIDENCEHYPOTHESISUNKNOWN

Supported by the dataset · a reasonable guess to test · not answerable from the data.

Your team marks each finding

KeepParkIgnore

Use it now · revisit later · drop it.

Evidence strength

STRONGMEDIUMWEAK

Consistent, multiple recent sources · some, limited or mixed · anecdotal or single source.

Your team setup

Optional. What you pick here is added to the top of every copied prompt as context, so Claude never loses track of your case. Saved in this browser only.