AK
Arunesh Kumar
Product Manager | Growth & Digital Strategy

9-Agent Product

9-Agent CRO Workflow

A structured multi-agent system for conversion optimization — breaking a business problem into nine specialist lenses, each contributing a distinct layer of product judgment.

The nine lenses

How the workflow thinks

Each agent represents a distinct layer of product judgment. The goal is not output volume — it is structured decision quality.

01

Data Agent

Establishes the baseline — what is actually happening in the funnel before any opinion enters the room.

02

UX Agent

Translates numbers into interface problems — where the experience breaks down and why users leave.

03

Research Agent

Grounds the diagnosis in market reality — are these problems unique to us or industry-wide patterns?

04

Customer Voice Agent

Brings the human layer — what buyers actually say, fear, and need before they convert.

05

Tech Agent

Stress-tests every idea against the stack — filters out what cannot be shipped before prioritization.

06

Experimentation Agent

Turns validated problems into testable bets — structured hypotheses with measurable success criteria.

07

Prioritization Agent

Forces a decision — which bets get resources first, ranked by impact, effort, and confidence.

08

Risk Agent

The final filter — what could go wrong, what are we assuming, and what needs a contingency.

09

Documentation Agent

Closes the loop — converts decisions into a format engineering can act on without a follow-up meeting.

Workflow

How it runs — step by step

Each agent hands off to the next, building a complete picture from raw data to a dev-ready execution plan.

Your business data

GA4 · Google Ads · Sheets dashboard

1 · Data Agent

Finds funnel leaks, drop-offs, and device splits using GA4, Google Ads, and dashboard data.

2 · UX Agent

Identifies page friction, confusing flows, and interaction barriers using data insights.

3 · Research Agent

Validates UX findings against market benchmarks and competitor behaviour.

4 · Customer Voice Agent

Maps findings to buyer psychology, trust gaps, and real customer language.

5 · Tech Agent

Validates technical feasibility on your stack and flags implementation constraints.

6 · Experimentation Agent

Designs A/B tests with hypothesis statements, variants, and lift targets.

7 · Prioritization Agent

Ranks every action by RICE and ICE score against effort and business impact.

8 · Risk Agent

Flags implementation risks, conversion risks, and edge cases before execution.

9 · Documentation Agent

Produces a dev-ready action plan, decision memo, and 30-day execution summary.

Final output

Top 3 actions · Revenue impact · 30-day plan

Color key

Data
UX + Research
Voice + Tech
Experiment
Risk + Docs

Three phases

1

Observe

Data Agent reads funnel signals, drop-offs, and device splits to frame the problem.

2

Diagnose

UX, Research, Customer Voice, and Tech agents each contribute a specialist lens on the problem.

3

Decide

Experimentation, Prioritization, Risk, and Documentation agents produce a ranked, dev-ready action plan.

Want to see the PM Hub?

The PM Hub contains 17 AI modules for end-to-end product work.

Open PM Hub