PM AI Agent Fleet: 39 Agents Mapped to the 7-Stage Operating System
A live index of every AI agent for product managers, mapped to the 7 stages of the PM Operating System: Sense, Discover, Decide, Build, Ship, Measure, Amplify.
The short version
The PM AI agent fleet is a set of autonomous agents that cover every repeated decision and output a product manager makes, mapped to the seven stages of the PM Operating System: Sense, Discover, Decide, Build, Ship, Measure, Amplify. Each agent runs on a schedule, pulls from connected data sources via Model Context Protocol (MCP), and delivers reports where you already work. Setup takes about an hour. Deploy them incrementally starting with Red Flag Detection. By month two the whole fleet is running. This page is the live index, auto-generated from the blog, and it updates every time I publish a new agent post.
For the honest meta-look at the fleet, what stuck, what died, what I'd do differently, see 39 PM AI Agents Deployed: What Stuck, What Died, and Why, the field-research piece behind this index.
Why a full fleet, and why full automation is a PM superpower
I started with 18 agents. They saved me 10+ hours a week. Then I realized the architecture was incomplete.
The agents were only covering parts of the PM job. I had good visibility into what was broken (Sense), but no agents orchestrating research into customer insights (Discover). I could spot priorities (Decide) but couldn't automatically generate PRDs or coordinate releases. The infrastructure was there, MCP wired up, vector DB running, but the playbook was thin.
So I mapped the entire PM Operating System across 7 stages and kept adding agents until each major decision or output a PM makes repeatedly had one. The result: I'm not doing PM work anymore. I'm doing PM thinking. The data flows, analysis happens, decisions get shaped, and I focus on judgment, strategy, and leadership.
This is what full PM automation looks like.
The data foundation
Your agent stack sits on three layers:
Structured data via MCP (Model Context Protocol) Connected data sources are everything. Agents pull from Jira, Slack, Zendesk, Salesforce, GitHub, Google Calendar, Notion, Google Drive, and analytics (Amplitude, Mixpanel). MCP standardizes the access, write once, agents read across all systems.
Unstructured data via vector database (Weaviate) Call transcripts from Gong. Support ticket descriptions. Meeting notes. Emails. These live in a vector database so agents can semantically search, finding patterns that share meaning, not just keywords.
Delivery and scheduling Agents output via Slack webhooks (reports land where you check them). They run on cron schedules (daily, weekly, bi-weekly). Automated but you control the cadence.
To get started:
- Claude setup guide, MCP connections for Claude Desktop, Claude Code, or Claude Projects
- OpenClaw setup guide, self-hosted alternative with the same data sources
Setup takes one hour. Then you have the foundation.
The 7-stage PM Operating System
Every agent I've written about slots into one of seven stages. The list below is generated from the blog at build time, so when I publish a new agent-*.mdx post and tag it with a stage, it appears here automatically. No article edit required.
Try the live sandbox first. Open the Agent Sandbox (PM Version) to poke at an interactive version of the fleet before reading every blueprint. You can click through agents, see sample outputs, and feel the cadence in about five minutes.
agent-*.mdxpost with agentStageand agentSchedulefrontmatter and it appears here automatically.SENSE, Signal Detection
6 agentsThese agents scan for what you need to know right now. Daily noise that might be signal.
Build Your Competitive Intelligence Agent
Bi-weeklyStop manual competitive research. This agent monitors competitors, tracks customer mentions, and delivers a weekly intel report so you always know your market.
Build Your Daily Red Flag Agent
Daily 9:00 AMThe first agent every PM should deploy: a morning scan of support, backlog, Slack, and production that tells you exactly what needs your attention today.
NPS and CSAT Deep Dive Agent
Daily 8:00 AM + Weekly MondayDaily and weekly analysis of your NPS and CSAT scores. Segment breakdowns, feature drivers, and what's actually making your customers happy or frustrated.
Automate Support Pattern Detection
Daily 8:00 AMA daily agent that extracts signals from your support queue: which problems are rising, which segments are hurting, and what real trend lies beneath the noise.
KPI Watchdog Agent: Catch Metric Drops and Ship a Fix Prototype
HourlyA watchdog agent monitors your KPIs hourly, investigates the root cause of any drop, and ships a working prototype fix before you finish your morning coffee.
Renewal Risk Agent for Migration Cohorts
DailyCatches the accounts whose operational dashboard says fine but whose qualitative signals say otherwise. Three-dimension scoring, 90 days before renewal.
DISCOVER, Research & Synthesis
5 agentsThese agents synthesize the noise into insight. Customer research, market trends, patterns, hypotheses.
Build Your Weekly Ops Digest Agent
Weekly Monday 8:00 AMDaily agents catch fires. This weekly digest spots the trends: recurring issues, degrading resolution times, and systemic problems that daily reports miss.
Build Your Market Intelligence Agent
Bi-weeklyBi-weekly deep dive into your competitive landscape. Win/loss analysis, market shifts, feature parity, and strategic recommendations so you're never surprised.
Customer Segmentation Agent
Weekly MondayWeekly updates to your customer segments and personas. Cohort analysis, segment evolution, and behavioral groupings - automatically maintained.
Customer Interview Synthesis Agent
Weekly WednesdayWeekly automated synthesis of your customer interviews. Key themes, hypotheses, and actionable insights - without the tedious manual work.
Automated Customer Journey Mapping
Bi-weekly ThursdayBuild rich customer journey maps bi-weekly from session replays, support patterns, and research data. See exactly where users get stuck.
DECIDE, Prioritization & Planning
8 agentsThese agents turn research into prioritized decisions. What to build? When? Why? This is where strategy becomes action.
Build Your Roadmap Progress Agent
Daily 9:00 AMYour roadmap says in progress but engineering hasn't touched it in weeks. This agent cross-references your backlog against actual dev activity daily.
Build Your Customer Commitment Agent
Bi-weeklySales promised a feature by Q2. CS said 'next month.' This agent tracks every commitment, cross-references the roadmap, and flags what is overdue.
Build Your Engineering Capacity Agent
Weekly Monday 8:00 AMEngineering capacity agent: flags overloaded engineers, catches PTO gaps weeks early, and tells you if the roadmap won't fit before the sprint starts.
Testable Assumptions Tracker Agent
Weekly FridayConvert opportunities into testable assumptions. Track validation status weekly. Know which assumptions are holding up your roadmap.
OKR Progress and Prediction Agent
Daily 4:00 PM + Weekly FridayDaily OKR tracking with outcome scoring. Weekly confidence predictions for OKR achievement. See what's on track and what needs intervention.
Opportunity Prioritization and Synthesis Agent
Weekly FridayWeekly synthesis of all your DISCOVER agent outputs into a prioritized opportunity stack. OST-ready opportunities with impact estimates and dependencies.
Automated Sprint Planning Agent
Weekly MondayConvert prioritized opportunities into user stories and sprint plans every Monday. Accounts for team capacity, tech debt, and dependencies.
Pricing Migration Tracker Agent
DailyWatches every migration account daily, classifies drift type, and recommends a specific intervention. Catch quiet drifters in week one, not week six.
BUILD, Prototyping & Execution
7 agentsThese agents keep the team unblocked and shipping. The day-to-day operations of building.
Build Your Product Ops Agent
Daily 9:00 AMDaily product ops report ranked by ARR at risk: top complaints, feature gaps, usage anomalies, and at-risk customers synthesized from five data streams.
Build Your Documentation Gap Agent
Daily 7:00 AMFeatures ship without docs and customers find the gaps. This agent scans tickets and the release calendar to catch documentation gaps before they become fires.
Build Your Team Triage Agent
Daily 9:00 AM and 4:00 PMYour #team-product channel is a firehose. This agent reads every message, categorizes issues, assigns owners, and surfaces what's unanswered, twice daily.
Automated PRD Generator
Weekly MondayConvert prioritized opportunities into PRDs automatically. Drafts based on research context, design specs, and technical requirements.
Tech Debt Impact and Prioritization Agent
Weekly MondayWeekly analysis of tech debt impact on velocity. Which debt items actually matter? What's the priority order to reclaim speed?
Auto Bugfix Agent: Zendesk Ticket to Reviewable PR in 11 Minutes
On-demand + Daily 8:00 AMAn AI agent that reproduces customer-reported bugs, locates the broken code, writes the fix, adds a regression test, and opens a reviewable PR.
Instant Prototype Agent: Customer Request to Prototype in Minutes
On-demand + Daily 9:00 AMAn AI agent that turns customer feature requests into working prototypes. Deploys a preview URL, opens a PR, files a Linear ticket, drafts a Notion doc.
SHIP, Release & GTM
5 agentsThese agents own the release process. From readiness to launch to post-release validation.
Build Your GTM Release Monitoring Agent
Daily 7:00 AMTrack if features are ready to sell, support, and succeed. Monitors discovery compliance, beta standards, and GTM materials daily across five readiness gates.
Build Your Release Readiness Agent
Weekly WednesdayPrevents launch disasters. Checks PRDs, release notes, feature flags, and GTM readiness every Wednesday so nothing ships without proper preparation.
Build Your Release Checker Agent
Weekly ThursdayThe final gate before you ship. This Thursday agent verifies QA, docs, GTM materials, and sign-offs, giving a clear go/no-go for each feature.
Automated Release Documentation Agent
Weekly Wednesday + per-releaseAuto-generate release notes, customer comms, and doc updates as features ship. Saves hours on documentation and keeps comms consistent.
Launch Comms Agent: Six Channels Generated on Every Release
On production deploy + Thursday 10:00 sweepAI agent that generates website copy, LinkedIn post, customer email, in-app banner, changelog, and X thread when a feature ships. Six channels, four minutes.
MEASURE, Impact & Health
8 agentsThese agents tell you if what you built actually worked. Did it move the needle?
Build Your Product Health Dashboard Agent
Weekly TuesdayGo deeper than daily metrics. Weekly agent analyzing feature adoption curves, cohort retention, engagement depth, and performance trends to shape strategy.
Build Your PM Issues Agent
Daily 7:00 AMCatch slipping features, overdue customer promises, and cross-team dependency failures before they become fires. Scans roadmap and commitments every morning.
Build Your Product Health Agent
Daily 4:00 PMDaily 4 PM pulse check: engagement, activation, performance, and support trends synthesized into one story. Know exactly where your product stands each day.
Feature Adoption Tracking Agent
Daily 4:00 PM + Weekly MondayDaily adoption curves for new features. Identify stuck cohorts and recommend interventions before adoption stalls.
Win/Loss Analysis Agent
Bi-weekly TuesdayBi-weekly analysis of won and lost deals. Extract product insights from sales outcomes. What's winning business, and what's holding you back?
Signal-to-Ship Cycle Time Agent: Measure PM Velocity Across 7 Stages
Daily 9:00 AM + Weekly MondayAn AI agent that tracks PM cycle time across the 7 stages of the Product Operating System (Sense to Amplify) and surfaces the weekly bottleneck.
Margin Watch Agent
Daily (operational) / Weekly (CPO digest)Tracks gross margin per outcome daily, classifies compression cause, and forecasts the Jevons cliff. Pricing reviews stop being post-mortems.
Direction Dashboard Agent
DailyCompiles seven leading indicators every morning, predicts outcomes 4-8 weeks ahead, runs a weekly Goodhart audit. Direction at AI-native velocity.
AMPLIFY, Communication & Learning
5 agentsThese agents share what you know with the people who need to know it. Strategy, leadership, alignment.
Build Your Executive Report Agent
Weekly Monday 1:00 PMAuto-generated Monday leadership brief: roadmap status, top wins, key risks, and metrics in a 3-minute read. Your CEO walks in already prepared.
Build Your Daily Focus Agent
Daily 7:00 AMYour AI chief of staff: reads your calendar, scans Slack, checks the roadmap, and delivers the 3 things that actually matter today before your first meeting.
Retrospective Synthesis and Learning Agent
Weekly FridayExtract learnings from sprint retros automatically. Update playbooks, surface patterns, and drive continuous improvement.
Stakeholder Communication Agent
Weekly Friday + MonthlyGenerate tailored updates for different audiences: exec summaries, board updates, investor reports, team briefings. All from the same data.
Board Narrative Drafter Agent
Weekly (drafts) / Quarterly (final)The most expensive document in product management compiled while you sleep. Quarterly board updates from data sources, comp set, and last quarter's narrative.
Your daily and weekly rhythm
Across the fleet, agents fire on a staggered schedule so information arrives when it's useful, not when it's fresh off the keyboard.
7:00 AM (daily, weekdays) Daily Focus, PM Issues, Documentation Gaps, GTM Monitoring fire. You know your priorities, what's broken, what's missing, and what's happening in market before you open email.
8:00 AM (daily, weekdays) Support Ticket Signals and NPS/CSAT Analysis run. Signal detection deepens, customer sentiment and issue patterns become visible.
Monday, 8:00 AM Weekly Ops Digest, Engineering Capacity, Customer Segmentation, and Sprint Planning kick off. This is your week-setup moment, patterns from last week, capacity reality, customer shifts, and your sprint crystallizes.
9:00 AM (daily, weekdays) Red Flag Detection, Product Ops, Roadmap Tracker, Team Triage, and Signal-to-Ship Cycle Time fire. Second wave of daily signals, catches what slipped through morning scan, and refreshes the in-flight portfolio view.
Monday, 9:00 AM Signal-to-Ship Cycle Time digest lands. The weekly meta-view of the whole fleet's output: where time is going across the seven stages (Signal, Prototype, Design partner, Production code, Rollout, Launch, Measure), which stage is the bottleneck, and how cycle time is trending. The agent that tells you whether the PM transformation is real.
Monday, 1:00 PM Executive Report compiles after morning agents finish. Prep for leadership conversations.
Monday, 4:00 PM Opportunity Prioritization and Assumption Tracker run. Your week's prioritized bets and validation plan are ready.
Tuesday, 9:00 AM Product Dashboard updates metric views before any stakeholder conversations.
Wednesday, 7:00 AM Interview Synthesis processes this week's customer calls into structured insights.
Wednesday, 8:00 AM Release Readiness Review and Release Documentation assess what's ready to ship.
Thursday, 10:00 AM Release Checker validates post-release stability.
Thursday, 2:00 PM Customer Journey Mapping updates your understanding of where friction lives.
Friday, 10:00 AM OKR Tracker and Retrospective Synthesis wrap the week. Did you hit outcomes? What did you learn?
Friday, 2:00 PM Opportunity Prioritization finalizes next week's top bets. Stakeholder Communication generates tailored updates for every team.
4:00 PM (daily, weekdays) Product Health and Feature Adoption run as your closing scan. OKR Tracker checks progress.
Hourly (continuous) KPI Watchdog watches your core metrics and pages you the moment a KPI drops outside its band, with a prototype fix attached.
Bi-weekly (rotating) Competitive Intelligence, Market Intelligence, Customer Commitments, Journey Mapping, Win/Loss Analysis cycle through. Strategic slower-burn agents that feed your monthly planning.
Building this yourself
The full agent fleet runs on:
- Claude (Desktop, Code, or Projects) with Model Context Protocol (MCP) connected to your data sources
- Weaviate vector database for semantic search across unstructured data
- Slack webhooks for report delivery
- Cron scheduling for agent execution
Setup guides walk you through all of it: Jira, Slack, Zendesk, Salesforce, GitHub, Google Calendar, Gong, analytics, PagerDuty, Sentry connections. One hour to wire. Then deploy agents incrementally.
Don't try to run them all at once. Start with SENSE (Red Flag Detection, Support Signals, NPS Analysis). Then add DISCOVER (Weekly Ops Digest, Interview Synthesis). Build from there.
By month two, your entire PM function is instrumented. By month three, you're operating at a completely different level.
What's still human
Agents handle data gathering, analysis, synthesis, and pattern detection. But PM judgment still lives here:
- Customer empathy. An agent tells you 14 customers mentioned a pain. Only you sit across from them and feel the real problem.
- Strategic trade-offs. The agent can prioritize by impact and effort. You decide whether market timing overrides the math.
- Organizational persuasion. Your agent can't convince your CEO that the roadmap shifted. You can.
- Hypothesis design. The agent flags assumptions. You decide which ones matter and how to test them.
- Creative problem-solving. The agent finds the constraint. You figure out how to design around it.
This is the PM job as it should always have been. Data gathering automated. Analysis automated. Thinking and judgment remain yours.
The honest limitations
They hallucinate. Claude is confident when it's wrong. Every output needs your eye. Always verify critical findings.
They need real data. A DISCOVER agent is useless if your Gong transcripts aren't in the vector DB. Garbage in, garbage out.
They need tuning. Deploy an agent with bad thresholds and it alerts you 20 times a day. Best agents are the ones you've trained over 2-3 weeks.
They miss context. An agent can tell you which customers churned. It can't explain the political situation that drove the churn.
They need management. Set up the foundation once. But manage the agents like a small team, review outputs weekly, adjust parameters, add new data sources, prune noise.
Start this week
Don't read about agents. Build them.
Today:
- Claude setup guide, wire MCP connections
- Start with Jira + Slack + Zendesk. One hour.
Tomorrow: Deploy Red Flag Detection and Daily Focus. These two reclaim 60+ minutes per day.
This week: Add Product Health and Support Signals. You now have basic signal detection.
Next week: Add Weekly Ops Digest, Interview Synthesis, and Market Intelligence. DISCOVER stage activates.
Week 3-4: Opportunity Prioritization, Assumption Tracker, Sprint Planning. DECIDE stage runs.
Month 2: Full fleet deployed and running. Your entire PM Operating System. You're operating at a different level entirely.
Your team will notice. Your manager will notice. Your customers will notice. That's the real point.
Frequently asked
What is the PM AI agent fleet?+
A set of autonomous AI agents that cover every repeated decision and output a product manager makes. The fleet is mapped to the seven stages of the PM Operating System: Sense, Discover, Decide, Build, Ship, Measure, Amplify. Each agent runs on a schedule (hourly, daily, weekly, or bi-weekly), pulls from connected data sources via MCP, and delivers reports to Slack or email.
What are the seven stages of the PM Operating System?+
Sense (signal detection), Discover (research and synthesis), Decide (prioritization and planning), Build (prototyping and execution), Ship (release and GTM), Measure (impact and health), Amplify (communication and learning). Every agent in the fleet maps to one of these stages.
How long does it take to set up the full agent fleet?+
One hour to wire the foundation (MCP connections to Jira, Slack, Zendesk, Salesforce, GitHub, Google Calendar, analytics, PagerDuty, Sentry). Then deploy agents incrementally, starting with Sense-stage agents (Red Flag Detection, Support Signal Processing, NPS Deep Dive). By month two, all agents are running. By month three, you're operating at a completely different level.
Which AI agent should a PM deploy first?+
Red Flag Detection. It runs daily at 9:00 AM, pulls from Zendesk, PagerDuty, Jira, Slack, and Salesforce, and scans for active incidents, unassigned tickets, escalations, and at-risk accounts. It reclaims about 45 minutes of morning triage per day and gives you a clean morning briefing before your first meeting.
What data sources do the agents need?+
Jira, Slack, Zendesk, Salesforce, GitHub, Google Calendar, Notion, Google Drive, analytics (Amplitude or Mixpanel), plus Gong or Chorus for customer call transcripts, and optionally Weaviate for semantic search across unstructured data. MCP (Model Context Protocol) standardizes access across all of them.
Do I need Claude Desktop, Claude Code, or something else to run the fleet?+
Any of three works: Claude Desktop with MCP, Claude Code (developer-focused), or Claude Projects. OpenClaw is a self-hosted alternative for teams that need on-premise data control. All four setups use the same agent definitions and MCP data connections.
What is the Signal-to-Ship Cycle Time Agent?+
The meta-agent of the fleet. It watches how every active product change moves through all seven PM OS stages, computes end-to-end cycle time, and names the weekly bottleneck. It tells you whether the PM transformation is real or performative by turning vague speed debates into a single compounding number.
Related reading
Deeper essays and other handbook chapters on the same thread.
The PM Agent Stack: Open-Source Tools Mapped to PM Work
Build a personal PM agent stack from open-source Claude repos. 18 tool categories mapped to the 7-stage PM operating system, with install order and how-tos.
The PM Agent Stack: A Bridge to the Enterprise AI Brain
Five-part series for PMs whose company is 6-18 months from an enterprise-wide AI brain. Build the personal PM agent stack today with open-source Claude repos.
39 PM AI Agents Deployed: What Stuck, What Died, and Why
An honest accounting of 39 PM AI agents across 4 product orgs in 80 days. Stage skew, cadence patterns, and the failure mode I kept repeating.
The 20-Minute Customer Call Triage Agent
Customer call triage agent: 60-minute call to tagged, theme-clustered artifact with a one-line action in 20 minutes. Stack, prompt, eval rubric, batch run.
Customer Discovery When Your Customer Is an Agent
Continuous discovery assumes a human customer. When your customer is an agent, six methods apply: agent telemetry, failure-mode interviews, capability audits.
Build a Discovery Agent Stack: Continuous Customer Listening
Discovery agent stack: seven open-source Claude repos turn customer interviews, support tickets, and public reviews into a continuous loop. Install + recipes.