Prototype improvement
Concierge Service Test
A Concierge Service Test simulates a real service by having people deliver the experience manually instead of relying on technology or automated systems. Depending on the setup, users may know they are interacting with a manual service (concierge), or the manual work may happen out of sight, creating the experience of automation. In both cases, the goal is to test the value, flow, and experience before investing in full implementation.
Why Use this tool?
Learn what truly needs to be automated, and what doesn’t.
This test helps you understand whether the service delivers value, how users experience it, and where friction appears. By running the service manually first, teams can learn faster, reduce risk, and make better decisions about what to build later.
what you should know
Start With: A service concept or experience with clear user interactions
End With: Learning about user value, service flow, and operational effort
Time Needed:
• Preparation: High
• Execution: 5–40 minutes
• Analysis: High
Difficulty: ⭐ ⭐⭐⭐⭐ (5 out of 5 – requires coordination and synthesis)
People:
• 8-20 users interacting with the service
• 2-5 team members delivering and observing the experience
A quickguide to start
2. Choose visibility. Decide whether users will see the people delivering the service or experience it as if it were automated.
3. Test internally first. Run a dry test to ensure the flow works.
4. Deliver to users. Fulfill real requests manually.
5. Observe closely. Track effort, timing, questions, and breakdowns.
6. Reflect and refine. Capture insights and adjust the service design.
helpful tips
- Both visible and behind-the-scenes setups are valid, choose based on what you want to learn.
- Focus on the experience and value, not speed or scale.
- Notice which steps feel heavy or repetitive, these often point to future automation opportunities.
RACU meets AI
Wizard of Oz
How Can AI Make RACU Easier ?
AI can be your creative partner and research assistant, ready to help you move faster and think deeper at every step of the RACU process.
For each RACU tool, we’ll share a ready-to-use AI prompt. Just copy the prompt into your favorite AI tool (like ChatGPT or Copilot) and it will guide you through the method step by step.
The AI becomes your facilitator, asking the right questions so you can build your thinking as you go. No need to fill out a blank form, the prompt starts the conversation and adapts to your answers in real time.
PROMPT – COPILOT, CHAT GPT
You are a facilitator helping me complete a Research & Discovery Card for a design thinking challenge.
Guide me step-by-step by asking the following questions one at a time, and wait for my answer before moving on. You can ask follow-up questions if needed to clarify or improve my responses.
Start with general context:
1. What is the challenge, project, or topic you’re working on? (Briefly describe the scope or goal.)
Then go into Research (existing data):
2. What existing information do we need to gather to better understand this challenge?
3. Where can we get that information? (e.g., internal reports, dashboards, previous research, public sources)
4. What specific questions will this data help us answer?
5. Who on the team will be responsible for gathering this information?
Then move to Discovery (new research):
6. Who should we learn from? (e.g., users, clients, collaborators, stakeholders)
7. Where can we find or reach them?
8. What topics, needs, or behaviors should we explore in the research?
9. What discovery methods could work best for this challenge? (Examples: interviews, shadowing, observation, journaling, immersing yourself in the experience, etc.)
10. How many people should we involve or study?
11. When will this research happen?
12. Who on the team will lead or coordinate this discovery work?
At the end, summarize my answers as a Research & Discovery Plan with two sections:
- Research (existing data)
- Discovery (new fieldwork)
Use bullet points and keep it simple enough to copy into a worksheet.
