Start with the customer experience and work backwards to the product.
Most product teams start with what they can build, or what a stakeholder wants, or a feature request.. and then find customers. Amazon's publicly documented approach inverts this: define the ideal customer experience first, write it down as a fictional press release, and only then decide what to build.
At Amazon, Patricia used the 5 Customer Questions (5CQ) for years. It's the first page of every cycle we run also in the AI-native Practitioner Programme, together with the PM AI Agent from the bot team. It's drafted before we write a PRFAQ. The job of the page is simple: understand the customer pain and the solution we want to build.
Each answer has to survive a quiet read.
Who is the customer?
One person, in a moment: time, place, situation. What insights do we have about this customer?
What is the customer problem or opportunity?
What is the pain or opportunity written from the customer's perspective? What data do we have to back it up?
What is the solution?
What is the solution this customer would keep? Why?
How does the customer experience look like?
Show it with sketches. Describe what is the most important benefit for this customer.
How will we define success?
What is the metric we will use to measure success? And its threshold?
Then the PRFAQ
When those five hold, interviews start. Then the PM writes the launch as if it already happened: a press release, then the FAQ a sceptical customer and a sceptical exec would ask. Agents may draft. The humans defend every line in a silent-read review.
The templates, the prompts, the uncomfortable question we would rather skip: that is the AI Native Practitioner Programme. A cross-functional team. A bar raiser. Seats and Cohort 1’s public record are on that page.
How the same page becomes the brief once agents are doing the typing: Working Backwards with agents.