Product management is fundamentally about making decisions with incomplete information. What to build, for whom, and when. AI doesn’t make those decisions for you—but it can dramatically improve the information you’re working with.

I’ve been a PM for eight years. The last eighteen months have changed my toolkit more than the previous eight combined.

User Research at Scale

Traditional user research is expensive and slow. AI is changing the economics.

Dovetail now uses AI to analyze interview transcripts, identify themes, and generate insights automatically. What used to require days of synthesis happens in hours.

Grain and Fathom record and transcribe user interviews, then use AI to identify key moments and generate summaries. You can search across all your interviews: “Show me every time a user mentioned frustration with onboarding.”

Hotjar and FullStory have AI features that surface interesting user behavior patterns without manual analysis. “Users who do X are 3x more likely to churn” emerges from the data.

For surveys, Typeform and SurveyMonkey use AI to analyze open-ended responses at scale. Hundreds of text responses get categorized and synthesized automatically.

ChatGPT or Claude can analyze raw feedback data you paste in. “Here are 50 customer support tickets. What are the main themes?” produces actionable synthesis in seconds.

Competitive Intelligence

Perplexity AI can quickly summarize competitor positioning, recent announcements, and market context. It’s not a replacement for deep competitive analysis, but it accelerates the research phase significantly.

Crayon and Klue track competitor changes automatically and use AI to surface what matters. Price changes, feature launches, positioning shifts—you’re notified rather than discovering them accidentally.

ChatGPT is surprisingly useful for analyzing competitor products. “Compare these two products’ feature sets and identify gaps” produces structured analysis quickly.

For market sizing, Claude or GPT-4 can help structure bottom-up and top-down estimates, identify assumptions, and sanity-check calculations.

Writing PRDs and Documentation

Product documentation is necessary but time-consuming.

Notion AI can generate first drafts of PRDs from your notes and requirements. “Turn these bullet points into a structured PRD with user stories” gets you 70% there.

Coda AI does similar work if you’re in that ecosystem. AI-generated tables, summaries, and formatted documents from raw input.

Claude is my go-to for longer documents. It handles nuance better than other models—important when you’re documenting complex requirements.

For release notes, AI-generated drafts from commit messages or Jira tickets save hours. You edit for voice and priority; the synthesis is automatic.

Loom with AI summaries lets you record quick video explanations and auto-generate written documentation. Engineers get the video; stakeholders get the text summary.

Roadmapping and Prioritization

Productboard now uses AI to analyze customer feedback and suggest features to prioritize. It connects feedback themes to potential roadmap items automatically.

Aha! has AI features for writing initiative summaries and generating roadmap presentations from your data.

For prioritization frameworks, Claude or ChatGPT can help structure RICE scores, weighted factors, and opportunity assessments. “Help me build a prioritization matrix for these 10 feature requests considering impact, effort, and strategic alignment.”

Linear with AI features can summarize sprint progress and predict delivery timelines based on historical velocity.

Stakeholder Communication

PMs spend enormous time translating between audiences. AI helps.

For executive updates: “Summarize this sprint’s progress in three bullet points suitable for a board deck” produces appropriate abstraction levels.

For engineering handoffs: Claude can translate product requirements into technical specifications, at least as starting points for engineering review.

For customer-facing content: AI drafts release announcements, feature documentation, and changelog entries.

Gamma and Beautiful.ai generate presentations from content. Quarterly business reviews, roadmap presentations, strategy decks—you focus on the message, AI handles the slides.

For email communication, Claude or ChatGPT helps draft stakeholder updates, difficult conversations, and diplomatic “no” messages for feature requests.

Analytics and Metrics

Mixpanel and Amplitude both have AI features that let you ask questions in natural language. “Show me retention for users who completed onboarding vs. those who didn’t” without writing queries.

Metabase with AI can generate SQL queries from descriptions. “Monthly active users broken down by plan type” produces the query automatically.

Obviously AI builds predictive models without data science expertise. “Which users are likely to upgrade?” becomes answerable with drag-and-drop data analysis.

For synthesizing metrics into insights, Claude handles complex data interpretation well. “Here are our key metrics from last quarter. What story do they tell?”

The PM Workflow Integration

Here’s how I integrate AI into my actual work:

Monday (planning): AI summarizes feedback from the past week, highlights urgent customer issues, drafts the week’s priorities

User research: AI transcribes and synthesizes interviews, identifies themes across multiple conversations

Writing: AI generates PRD first drafts, release notes, stakeholder updates—I edit and refine

Meetings: AI takes notes, generates summaries, tracks action items

Analysis: AI helps interpret data, build models, identify patterns

Communication: AI drafts emails, presentations, documentation

The human work: Deciding what matters, making tradeoff judgments, building relationships, understanding context that data doesn’t capture.

What AI Can’t Do

Strategic judgment. What should the product become? AI can analyze options but can’t feel the vision.

Stakeholder politics. Understanding organizational dynamics, building coalitions, navigating conflicting interests—entirely human.

Customer empathy. The deep understanding of user problems that drives great products comes from human connection, not data analysis.

Saying no. Prioritization requires disappointing people. AI can inform the decision; you have to own it.

Accountability. When things go wrong—and they will—you’re responsible. AI doesn’t bear consequences.

The PM’s AI Stack

Essential:

• Claude Pro ($20/month): Writing, analysis, synthesis
• Grain or Fathom (free tiers available): Interview recording and analysis
• Notion AI ($10/month): Documentation
• Gamma (free tier): Presentations

Worth the investment:

• Dovetail (team pricing): Scaled user research
• Productboard (team pricing): AI-enhanced feedback analysis
• Mixpanel/Amplitude (team pricing): Natural language analytics

The Multiplier Effect

The best PMs I know aren’t using AI to do less. They’re using it to do more.

More customer conversations (because synthesis is faster). More experimentation (because documentation is cheaper). More strategic thinking (because administrative work is automated).

AI amplifies PM capability. A PM with strong AI skills now produces what two or three PMs did before—not by working harder, but by leveraging tools effectively.

That’s the competitive reality. The PMs who embrace these tools become dramatically more effective. The ones who don’t become relatively less valuable.

The product management skillset is evolving. AI literacy is no longer optional—it’s core to the role.

---

PM tools are evolving rapidly. I’ll update this as better options emerge.