Skip to content
Module
Back to courses
Product ManagementAdvanced

Set up continuous discovery with AI

A weekly interview routine, AI analysis of customer feedback, traceable synthesis and an opportunity solution tree kept up to date.

Course designed by Alex Dionisio, Senior Product Manager. Updated November 14, 2026.

  • 59 steps
  • ~3 hr 4 min

Level: Advanced · Regular practice: you already use the tools or work on this topic.

Know how to set up a discovery practice that runs every week and make use of AI without losing touch with what your customers actually say. You learn to get interviews into your calendar without recruiting effort, to record and transcribe them in line with consent and the GDPR, to analyze tickets, reviews, NPS comments and calls at scale with an LLM, a taxonomy and a validation sample, to keep a repository where every insight traces back to its evidence, to spot fabricated quotes and overgeneralizations in an AI-generated synthesis, to keep your opportunity solution tree and your assumptions alive, and to decide when synthetic users help and when they mislead. You leave with the continuous discovery kit for your product: the routine set in your calendar, the consent kit, the feedback analysis pipeline, a seeded repository and an up-to-date tree.

What you will be able to do

  1. Set up a weekly interview routine, with continuous recruiting and clear roles within the product trio.
  2. Record and transcribe interviews with AI tools while respecting consent and the GDPR, then check the transcript.
  3. Analyze existing feedback (tickets, reviews, NPS, calls) at scale with an LLM, using a taxonomy, structured outputs and a validation sample.
  4. Build an insight repository where each insight is linked to its evidence, and share it with the team along with its confidence level.
  5. Check the fidelity of an AI-generated synthesis and correct fabricated quotes and overgeneralizations.
  6. Keep an opportunity solution tree up to date, map risky assumptions across five categories and choose the right test for each.
  7. Decide when synthetic users can be useful and when they mislead.

Prerequisites

  • Work in a product role (Product Manager, product designer or product owner) on a product in production
  • Use a conversational AI assistant routinely; no technical skills are required
  • Recommended course before this one: Design and validate a product experience with AI, which covers running interviews, a first opportunity solution tree and personas.
  • To write good synthesis prompts, see also the course Prompting for Product Managers: delegate, describe, verify.

Syllabus

What will I be able to do by the end of this course, and in what order?

  1. Objective · By the end of this overview, you will know what you are going to produce, in what order, and how Atelio's continuous discovery will serve as the running example.

How do you fit conversations with users into every week, without a research project or recruiting effort?

  1. Objective · By the end of this lesson, you will be able to explain why a weekly interview cadence teaches you more than a one-off study, and to split the trio's roles in that routine.

  2. Objective · By the end of this lesson, you will be able to set up a recruiting channel that fills your slots every week, with a screener question and a buffer for no-shows.

  3. Objective · By the end of this lesson, you will be able to prepare a compliant consent kit, choose a transcription tool with full knowledge of the trade-offs and check a transcript against the audio.

How do you get a reliable reading out of hundreds of tickets, reviews and comments, and know what it does not tell you?

  1. Objective · By the end of this lesson, you will be able to inventory your customer feedback sources and say, for each one, what it reveals and what it does not show.

  2. Objective · By the end of this lesson, you will be able to write a theme taxonomy and an output schema, then have an LLM code a batch of feedback in a way that can be counted and checked.

  3. Objective · By the end of this lesson, you will be able to measure the quality of an LLM's coding on a sample reviewed by humans and fix the taxonomy where the model gets it wrong.

How do you make sure every insight shared with the team rests on verifiable evidence, even when AI wrote the synthesis?

  1. Objective · By the end of this lesson, you will be able to organize a research repository where each insight is linked to the quotes, tickets and reviews that support it.

  2. Objective · By the end of this lesson, you will be able to review an AI-generated interview synthesis, spot fabricated quotes, overgeneralizations and misreadings in it, and correct them.

  3. Objective · By the end of this lesson, you will be able to share insights with the team along with their confidence level and the link to their evidence, without turning a lead into a certainty.

How do you evolve your opportunities and assumptions at the pace of what you learn each week?

  1. Objective · By the end of this lesson, you will be able to evolve an opportunity solution tree based on the latest interviews, map a solution's assumptions across five categories and choose a test for the riskiest ones.

  2. Objective · By the end of this lesson, you will be able to decide in which cases synthetic users can help your discovery and in which cases they lead it astray.

What does a continuous discovery system on your product look like, ready to run as early as next week?

  1. Objective · By the end of this workshop, you will be able to assemble the continuous discovery kit for your product and assess it with a grid covering each course objective.

Glossary terms

Public definitions of the key concepts in this course:

Full glossary