What can AI automate for a Growth PM?

AI can help with: Drafting multiple copy and creative variants for tests; Pulling cohort, funnel, and retention data into readable summaries; Writing first-draft experiment recap and PRD docs; Tagging and clustering qualitative user feedback; Calculating sample sizes and basic significance checks.

What stays distinctly human for a Growth PM?

Still human: Deciding which growth bets actually matter to the business; Judging whether a result is real or a fluke worth retesting; Building trust with engineering, design, and data partners; Choosing what not to test when resources are limited; Setting the narrative for leadership on what the team learned.

AI for Growth PM: tools, prompts, and how the role is changing

The shift

How AI is changing the Growth PM role

In 2026, AI handles much of the grunt work behind growth experiments: writing test variants, summarizing user research, and pulling cohort data into plain-language readouts. Growth PMs now spend less time wrangling spreadsheets and more time deciding which bets to make and why. The role is shifting toward judgment about where to test and how to interpret messy results, not the manual work of setting tests up.

What AI can take off your plate

Drafting multiple copy and creative variants for tests
Pulling cohort, funnel, and retention data into readable summaries
Writing first-draft experiment recap and PRD docs
Tagging and clustering qualitative user feedback
Calculating sample sizes and basic significance checks

What stays distinctly human

Deciding which growth bets actually matter to the business
Judging whether a result is real or a fluke worth retesting
Building trust with engineering, design, and data partners
Choosing what not to test when resources are limited
Setting the narrative for leadership on what the team learned

Tools

Five AI tools for Growth PMs

Amplitude

A Growth PM uses its AI query and chart features to ask funnel and retention questions in plain English and get cohort breakdowns without writing SQL.

ChatGPT

A Growth PM drafts onboarding copy variants, experiment hypotheses, and PRD outlines, then refines them against past test results.

Mixpanel

A Growth PM uses Spark, its AI analyst, to investigate drop-off points and surface which user segments convert differently.

Notion AI

A Growth PM summarizes experiment notes, generates test recap docs, and keeps a searchable backlog of past learnings.

Statsig

A Growth PM runs and reads A/B tests, with AI help interpreting significance and flagging which metrics moved.

Prompts

Five prompts to try today

Paste these into Claude or ChatGPT and replace the bracketed parts with your own details.

1. Generate experiment hypotheses

Our [funnel step, e.g. signup to activation] has a [X]% drop-off. Here is what we know about the audience: [context]. Give me 8 testable hypotheses, each with a clear expected mechanism and the metric it should move.

2. Write onboarding copy variants

Write 5 variants of [screen or email] copy for [product] aimed at [user segment]. Goal is to increase [metric]. Keep each under [word count] words and vary the angle: clarity, urgency, social proof, value, and curiosity.

3. Read a test result

Here are A/B test results: control [numbers], variant [numbers], sample size [N], duration [days]. Explain in plain language whether this is conclusive, what the likely effect size is, and what I should do next.

4. Prioritize the backlog

Here is my experiment backlog with rough impact, confidence, and effort estimates: [list]. Score each with ICE, explain the top 3, and flag any I should drop or combine.

5. Summarize user research

Here are [N] user interview notes about [topic]: [paste]. Pull out the top recurring frustrations, surprising quotes, and 3 growth opportunities tied to specific funnel steps.

A day in your inbox

This is the kind of brief a Growth PM gets, every weekday morning.

The Morning Current

Weekday morning

✦ Personalized for: Growth PM

Today's Tool

Find the leak with Mixpanel Spark

Ask Spark to compare conversion across your last three signup cohorts and break it down by acquisition source. It often surfaces a segment-specific drop you would have missed in an aggregate chart.

Today's Prompt

Turn the leak into hypotheses

Paste the cohort data into ChatGPT and use the experiment hypotheses prompt. You will get a ranked list of testable ideas tied to the exact funnel step that is leaking.

Today's Trick

Pre-commit to a decision rule

Before launching, write down the metric, the threshold, and what you will do at each outcome. This stops you from rationalizing a flat result after the fact and keeps tests honest.

AI for Growth PMs