What is generative AI? A guide for product managers
Generative AI refers to artificial intelligence systems that create new content, whether text, images, code, audio or video, from an instruction. It runs on models trained on huge volumes of data, such as large language models for text or diffusion models for images. It differs from predictive AI, which classifies or estimates a value without producing new content.
Definition written by Alex Dionisio, Senior Product Manager. Updated November 13, 2026.
Why it matters for a PM
Generative AI has opened up features that used to be out of reach for small teams: assisted writing, summaries, natural language search, assistants, image creation. It also brings new problems for a PM: variable answers, mistakes stated with confidence, a cost on every use, questions about rights and data. Framing it well is now part of the job.
Example
A peer-to-peer marketplace adds a "write my listing" button. From three photos and a few words, a model drafts a title and a description, which the seller reviews and edits before publishing. The PM defines what the text must never claim, such as the condition of something not visible in the photos, and tracks the share of listings published without edits.
Key points
- It produces likely content, not verified content, so review or checks remain necessary depending on the stakes.
- Its strongest product uses are the ones where a mistake is visible and easy to fix.
- Every generation has a cost and a delay that belong in the business model.
- It raises questions of transparency, copyright and data protection that are best handled early.
Common mistakes
- Adding a generative feature because it is expected, without a specific user problem to solve.
- Measuring success by clicks on the button rather than by the quality of what users keep.
- Presenting generated content as reliable information without labeling it.
Go further with Module
The courses and lessons that cover this concept:
Guides on this topic
- Claude Skills: build and test a skill
SKILL.md anatomy, install, triggering, testing and security, with a template.
- RAG vs fine-tuning: how to choose
What each approach solves, the decision criteria and how to evaluate, with a decision grid.
- Evaluating an AI feature (evals)
Test cases, criteria, LLM-as-a-judge, release thresholds and monitoring, with an evals plan.