Use a permission aware internal agent to scan customer feedback across support, calls, product systems, and shared company sources, then narrow from broad themes to cited evidence that a product team can inspect.
Source episode · 12:43
How I AI: Amplitude's Viral Internal AI Tool for Product Developmentwith Claire Vo
Wade starts with a broad query across product feedback, narrows into a theme such as linking session replay to funnel analysis, and opens the cited Zendesk and customer call evidence to decide whether the apparent demand has real heat behind it.
A theme report that quantifies coverage, links every finding to inspectable evidence, preserves contradictory feedback, and distinguishes customer language from the model's interpretation.
Step by step
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
4 steps
Connect only approved enterprise sources and preserve their permissions. Normalize source, timestamp, account or segment, product area, record type, and durable citation without flattening all feedback into one anonymous text pool.
Begin with a bounded question and require the agent to disclose its sample before listing themes. Ask for frequency, distinct sources and accounts, representative evidence, counterevidence, and confidence.
Analyze feedback about [product area] from [date range] for [segments]. First report accessible sources, exclusions, record counts, and sampling limits. Then list themes with frequency, distinct account count, exact cited excerpts, counterexamples, and confidence. Do not invent quotations or infer demand beyond the evidence.
Choose one theme and retrieve the underlying records. Compare what customers were trying to do, where the friction appeared, who was affected, and whether the same words describe different jobs or problems.
Investigate the theme [theme]. Group evidence by customer job, segment, workflow stage, and observed consequence. Link every claim to the original record, include contradictory evidence, and identify questions that require customer follow up.
Turn the evidence into decisions, hypotheses, and research gaps. Keep exact customer language beside the synthesis, and let product owners inspect the cited records before using the result in prioritization or a PRD.
From the previous analysis on [connecting session replay to funnel analysis], extract the 5 most representative customer quotes. For each quote, include the source (e.g., Zendesk ticket, Slack message) and a one-sentence summary of the user's core problem. Format the output as a bulleted list.
Turn an idea into a PRD, user stories, and a plan.
After the steps
How to recover when the loop fails and where human judgment helps.
Recover
Keep building
Design a weekly meta-workflow that analyzes your interactions with your AI assistant to make it smarter. This system learns your writing style from edits, suggests new automations for repetitive tasks, and helps you filter valuable AI techniques from hype.
Create a daily AI workflow that summarizes yesterday's meetings, identifies urgent tasks, and proactively learns internal company terms it doesn't understand. This keeps your AI assistant's knowledge current and highly relevant to your work.
Build a recurring AI task that automatically gathers updates from your calendar, Slack, and notes to suggest weekly priorities and prepare you for upcoming meetings. Start every Monday with a clear, comprehensive plan without the manual effort.
Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.