Code feedback at scale with an LLM
Handing hundreds of tickets or reviews to a language model quickly yields an appealing summary that nobody can verify. This lesson shows how to frame the analysis so the results can be counted, filtered and reread. You will know how to prepare a first batch of customer feedback and launch coding you can repeat.
Updated November 14, 2026.
In short
To analyze customer feedback reliably with an LLM, you give it a closed list of themes defined in advance and require a precise response format, with a source quote backing every label. For a PM, the output can then be counted, filtered and reviewed, instead of a summary that cannot be verified.
Lesson 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.
Questions this lesson answers
How do you analyze support tickets with an LLM?
Prepare a theme taxonomy with definitions, ask the model to label each ticket in a required format, then review a sample before relying on the numbers.
Why use structured outputs to classify customer feedback?
Binding the reply to a schema means every label comes from a permitted list, so results add up cleanly and anything off-frame is easy to spot.
Topics covered
- analyze support tickets with an LLM
- AI thematic coding
- LLM structured outputs
- analyze app store reviews with AI
- customer feedback taxonomy
In the glossary
Where it fits
Use what your customers already tell you
How do you get a reliable reading out of hundreds of tickets, reviews and comments, and know what it does not tell you?
Lessons in this module
- Take stock of what your customers already tell you
- Code feedback at scale with an LLM (this lesson)
- Check an LLM's coding on a sample
What you will learn in the course
This lesson is part of the course Set up continuous discovery with AI
Related courses
- AI & ProductAdvancedRegular practice: you already use the tools or work on this topic.
Get reliable, product-ready AI outputs (JSON, schemas, function calling)
2 hr 30 min
- AI & ProductAdvancedRegular practice: you already use the tools or work on this topic.
Build an AI assistant for your product
3 hr
- AI & BusinessAll levelsNo prerequisites: accessible to everyone.
Put AI to work in your business
2 hr 10 min