Evaluate product ideas against business metrics



Year: 2026 | Industry: Fintech | Region: Japan |  Role: UX Strategist/Prompt engineer | Team: UX Research | Deliverable: Business Impact GPT (Chat GPT)
  • Fintech is a notoriously business‑driven industry where metrics are painstakingly tracked. Designers know these metrics but often lack tools to assess whether their product ideas will deliver the expected business impact, especially when facing stakeholders who demand strong justification.

  • Together with the UX Research team, we identified three points in the design process where this issue arises:
1. During the conceptual and design iteration phase.
2. Before the first stakeholder review.
3. Before the final review with stakeholders.

  • The solution comes in the form of a business‑focused GPT that evaluates product concepts, strategies, and even rough ideas against business metrics such as long‑term value, retention, and GMV impact. To avoid “drift” or “contextual overload,” the GPT does not review aesthetics or delight. It only answers whether the proposed system increases transactions.
  • Designers and project managers can paste or upload documents into the chat. The GPT will then use established business standards to deliver a structured, detailed evaluation, even if the document includes little or no explanation.

  • To prioritize speed, the GPT evaluates uploaded documents using quarterly‑updated company datasets and patterns instead of real‑time data sources. The GPT evaluation includes:
1. Business verdict
2. GMV impact score
3. Customer short-term transactional behavior
4. Customer long-term transactional behavior
5. Risks
7. Measurement plan
8. Next steps

  • Because most people on our 40+‑person design team are non‑native English speakers, the final evaluation uses plain English with minimal business jargon and is formatted for easy copy‑pasting into slide decks or Slack.

  • The results show repeat usage among a core group of designers, but the three most important achievements are:
1. Empowering designers.
2. Eliminating personal-opinion based preferences.
3. Opening up discussion about the feasibility of certain product proposals.

Special thanks to the UX research team: Beau Tana, Pallavi Varma, Jane Susiriwatananont.