Definition
Team Swachh's idea asks residents to photograph bins. What if a photo shows a person's face, or someone's house number? What if some residents have no smartphone? A good idea can still cause harm if nobody thinks about these questions.
This step identifies everyone affected by your solution, the benefits each group could gain and the risks each could face. It combines a stakeholder map, user research and a risk assessment, so that the design serves real users and reduces harm before anything is built.
In simple words
Before building a new path through a park, you would ask who will walk on it, how it will help them and who might trip or be left out: people with wheelchairs, children, older people. This step asks those questions about your AI idea.
Why it matters
AI systems often help one group while quietly harming another. Mapping users, benefits and risks early shows you what to design for, what to avoid and whom to ask, and makes your final proposal far more convincing.
Key terms
- User
- A person who directly uses the solution.
- Indirect stakeholder
- Someone affected by a solution without using it.
- User persona
- A short description of a typical user, based on research, to keep their needs in mind.
- Benefit
- A positive effect the solution could have for someone.
- Risk
- A possible harm, described by how likely it is and how serious it would be.
- Mitigation
- An action that reduces a risk.
Explained in detail
Who is affected
Team Swachh drew a stakeholder map. At the centre: residents and shopkeepers who would send photos, and the sanitation supervisor and workers who would use the daily map. Further out: people living near the bins, stray animals affected by plastic, the municipal council that pays for trucks, and people who might appear in photos by accident. They also asked who might be left out: residents without smartphones, older people unfamiliar with apps, and people who do not read English.
User personas
From their interviews they wrote two short personas. Ramesh, 52, runs a vegetable stall, has a basic smartphone, reads Hindi, and is busy all morning; he will send a photo only if it takes a few seconds. Sunita, 38, is a sanitation supervisor who plans routes on paper each evening; she needs a simple list or map she can trust, not a complicated dashboard. Writing personas kept the team designing for real people instead of imagined ones.
Benefits and risks
| Group | Benefit | Risk | Mitigation |
|---|---|---|---|
| Residents | Cleaner streets | Photos might capture faces or homes | Ask users to photograph only the bin; blur or reject photos with people |
| Sanitation team | Less wasted time | Wrong predictions send trucks to the wrong bins | Show confidence; supervisor makes the final decision |
| People without smartphones | None directly | Their areas may be under-reported | Add a phone-call or shopkeeper reporting option |
| Workers | Better-planned routes | Being judged unfairly by the app's numbers | Use the map for planning, not for punishing workers |
Rating risks
Not all risks are equal. For each one, ask how likely it is (low, medium, high) and how serious it would be (minor, moderate, severe). A risk that is both likely and severe must be dealt with before going further; an unlikely, minor risk can be noted and watched. Team Swachh rated "photos capturing faces" as likely and moderately serious, so they built the mitigation into their design from the start rather than adding it later.
Finally, the team took their stakeholder map and risk table back to two of the people they had interviewed. The shopkeeper pointed out a risk they had missed: if the app showed which bins were full, some people might dump extra garbage at the emptiest bins near their own shops, moving the problem rather than solving it. The team added it to the table and planned to watch for it in any pilot. Checking your thinking with real stakeholders almost always reveals something new.
How it works
- Draw a stakeholder map with users at the centre and others around them.
- Write one or two user personas from your interviews.
- For each group, list benefits and risks.
- Rate each risk by likelihood and seriousness.
- Write a mitigation for every serious risk.
Try it yourself
The face in the photo
Team Swachh tests their idea by asking friends to send photos of bins.
Match the risk to a mitigation
How could each risk be reduced?
- Photos capture people's faces
- Residents without smartphones are left out
- Wrong predictions waste truck trips
- Workers judged unfairly by the numbers
Common mistakes
- Thinking only about users and ignoring indirect stakeholders.
- Listing risks without rating them or planning mitigations.
- Designing for imagined users instead of real ones.
Check yourself
Summary
- Map users and indirect stakeholders, and write personas from real research.
- List benefits and risks for each group; rate risks by likelihood and seriousness.
- Build mitigations into the design from the start.
Finished reading?
Sources: IDEO U: What is design thinking? · UNESCO Recommendation on the Ethics of Artificial Intelligence