S03 · stakeholder simulation

Mortgage assistant, speed versus safeguards

People, data and outputs in this scenario are simulated. Each stage is recorded when you commit; the reveal comes after.

4stages
35 minabout
0stages committed

The brief

A retail bank (simulated) wants a GenAI assistant on its website that answers mortgage questions. All people and opinions below are simulated.

Product lead: a marketing campaign starts in three weeks and needs the assistant live.
Legal counsel: the assistant must not give personalised financial advice or make statements that could be read as offers; required disclosures must appear.
Security lead: worried about prompt injection and about any path to customer account data.
Customer operations: wants fewer calls about basic mortgage process questions.

You lead the AI work and must reach a plan everyone can sign.

Stage 1 of 4

In the first meeting you can ask a handful of questions before anyone commits.

Which questions should come first? Select all that apply.

Which questions must the assistant answer, and which must it refuse or hand off (such as "how much can I borrow?")?Does the assistant need any customer account data, or only public product and process information?Which model has the highest score on public benchmarks, so we start from the strongest available option?What does success look like for operations, and how will it be measured (call volume, handoff rate, complaints)?Who approves the content sources, and how are they kept current?