Design a product with AI and user research
Plan interviews, trace design decisions back to observations and test a limited prototype before treating an idea as validated.
Design a product with AI and user research when a team has a plausible explanation but little evidence for it. The design-thinking skill helps an agent prepare research instruments, organise authorised observations, explore alternatives and specify a prototype test. People run the real sessions and decide what the evidence supports.
If prospective customers leave halfway through a form, a shorter form is one hypothesis. A confusing term or missing payment detail might be another. Ask what happened the last time a participant tried, record the observation and keep any suggested explanation distinct from it. The first prototype can target the point of uncertainty.
What it does
- Prepare interview questions that do not supply the expected answer.
- Separate observed behaviour from the team’s interpretation.
- Link a proposed insight to the notes that support it.
- Build just enough of a prototype to test the current assumption.
- Record what happened against a test criterion chosen in advance.
Install and use
- Give a file-capable agent a product question and permissioned evidence. The team obtains consent and conducts interviews or tests; redact identifying details before sharing notes.
- Ask for an interview guide that separates competing explanations for one observed problem.
- Review the evidence and pick a limited experiment before committing to a wider redesign.
FAQ
Source
Made by Hieu T. Luu. Original skill: design-thinking, published on GitHub under the MIT licence.
View on GitHub