The Product Owner in the AI Era

AI can greatly reduce the time a PO spends writing documents, but it does not assume responsibility for the product.

Tasks Well Suited to AI

Summarizing interviews
Comparing competing products
Clustering feedback
Drafting requirements
Proposing acceptance criteria
Exploring edge cases
Organizing tickets
Drafting release notes
Finding anomalous metric patterns
Summarizing meetings and extracting decisions
Generating prototypes / mockups
Investigating the codebase

Responsibilities Humans Must Retain

Which users to prioritize
Which problems to solve
Which trade-offs to accept
Which evidence to trust
How far to reduce scope
Final quality standards
Product direction

An AI-Assisted PO Workflow

flowchart TD
    H[Human PO] --> A[AI Research / Synthesis]
    A --> O[Options]
    O --> H
    H --> D[Decision]
    D --> W[AI-assisted Specification]
    W --> E[Engineering]
    E --> R[AI Review / QA Support]
    R --> H

Example AI Agent Roles

Agent	PO Tasks
Researcher	Investigate feedback, competing products, and code
Analyst	Analyze metrics and patterns
Spec Writer	Draft feature specifications
Reviewer	Review missed edge cases
Technical Explorer	Investigate areas affected by implementation
Meeting Scribe	Extract decisions and actions

Principles for Effective Use

1. Make the Decision Context Explicit to AI

Poor instruction:
"Plan a feature."

Good instruction:
"The goal is X, the users are Y, and the constraints are Z.
Compare 3 alternatives and show the trade-offs of each.
I will make the decision."

2. Connect It to the Codebase and Real Data

To prevent AI from offering only generalities, provide real context:

The actual repository
Actual user feedback
Actual metrics
Actual designs
Actual issues

3. Include an Agent That Challenges the Proposal

Do not rely only on agents that advocate for a feature.

Proposal
↓
Adversarial Review
- Is it really necessary?
- Can existing features solve it?
- Is the scope too large?
- Does it conflict with the user's mental model?

PO Skills That Become More Important in the AI Era

These competencies become more important, rather than less important, because of AI:

Problem Framing
Taste / Product Judgment
Prioritization
Systems Thinking
Technical Literacy
Evidence Evaluation
Decision Ownership

As AI lowers implementation costs, the relative cost of choosing what to build grows.

In the AI era, a PO should therefore be someone who repeatedly makes better choices in an environment that enables rapid experimentation, rather than someone who produces more documents.