The AI-Powered Weekly Review: A System for PMs Who Ship

The short version

The PM weekly review takes 30 minutes instead of 3 hours when you let AI agents pull the data, surface insights, and prepare a brief. Three phases: gather (10 minutes, mostly automated, pull from GitHub, Jira, analytics, Zendesk, Salesforce), triage (10 minutes, five decision questions about what changes your week), communicate (10 minutes, generate the standup opener for your team). Run a 4-week trend prompt at the end of each month to catch slow-moving problems that look fine in isolation. The system works because it's regular, surfaces surprises, creates 30-minute constraints, gets better through repetition, and frees you for what only PMs can do: customer conversations, strategic thinking, unblocking teams. Start with manual data collection. Automate when it works.

The PM's weekly review is where you either gain control of your week or let it control you. The old way: spend 2-3 hours on Monday morning digging through Slack, JIRA, analytics dashboards, and email to figure out what actually happened last week and what should happen this week.

The new way: Have AI agents pull the data, surface the insights, and prepare a brief you can review in 10 minutes. Then spend your remaining 20 minutes on what matters: deciding what ships, unblocking your team, and talking to customers.

This is the exact system I use and have refined with a dozen product teams. It's tactical enough that you can implement it today.

The Three-Phase Weekly Review (30 Minutes Total)

Phase 1: Gather the Brief (10 Minutes, Mostly Automated)

The goal: Get a one-page summary of what happened last week without reading 100+ messages.

Step 1: Set up your data sources (before Monday)

For this to work, you need to aggregate data from:

• Engineering: What shipped, what's in progress, what's blocked (from GitHub, Jira, or Linear)
• Product metrics: Key changes in activation, retention, DAU, conversion (from your analytics tool)
• Customer feedback: Top issues, feature requests, churn signals (from support tickets or review aggregator)
• Sales: Deals won/lost, objections, pipeline changes (from Salesforce or Pipedrive)
• Marketing: Campaign performance, website traffic, content metrics (from Segment or Google Analytics)

The setup time: 30-45 minutes, one-time. Automate this with Zapier, Make, or a simple script that feeds data to a shared doc or spreadsheet on Friday afternoon.

Step 2: Generate the AI brief (Monday 8am)

Use this prompt in Claude or your AI tool of choice:

I'm a Product Manager doing my weekly review. Here's what happened last week across our product:

ENGINEERING:
[Paste: what shipped, what's in progress, blockers]

PRODUCT METRICS:
[Paste: DAU/MAU change, activation rate, churn, key feature adoption]

CUSTOMER FEEDBACK:
[Paste: top 3-5 support tickets, any churn signals, feature requests]

SALES:
[Paste: deals won (with reasons), deals lost (with objections), pipeline movement]

MARKETING:
[Paste: top performing content/campaigns, traffic sources, key metrics]

Generate a weekly brief that:

1. **Headline**: One sentence on the most important thing that happened
2. **What Shipped**: Features/fixes from last week, metrics if available
3. **What's Momentum**: Metrics moving in right direction (growing, improving)
4. **What's at Risk**: Metrics declining, customer issues escalating, team blockers
5. **Key Insights**: 2-3 patterns or anomalies (e.g., "activation dropped 8% but retention is flat" or "we're hearing price objections from one segment")
6. **This Week's Context**: Anything external (competitor news, holiday, planned downtime) that matters

Format: Scannable bullets, no paragraphs. Max 400 words.

This takes 2-3 minutes to run. You now have a one-page brief instead of skimming through 50 Slack messages.

Pro tip: Include last week's brief alongside this week's. The AI will surface week-over-week trends automatically.

Phase 2: Triage and Decide (10 Minutes)

Now that you have the brief, you're deciding: What matters? What changes our roadmap? What do I need to talk to the team about?

Use this triage checklist:

Questions to ask of the brief:

1. Is anything broken that wasn't yesterday? (Metric dropped hard, customer issue escalated, deal lost)

   • If yes: This moves to top priority, might shift the week
   • Action: 10-minute sync with the relevant owner (eng, sales, support) to understand root cause and response

2. Did we learn something that changes our roadmap? (Unexpected usage pattern, new customer segment demand, competitor move)

   • If yes: Does this change what we ship this quarter, or is it a note for next quarter?
   • Action: Flag for roadmap review meeting

3. Are we hitting our goals? (On track, behind, ahead on quarterly OKRs)

   • If yes: Keep momentum, any quick tweaks?
   • If no: Do we need to shift what we ship, or did we just get unlucky one week?
   • Action: Escalate if a trend, not if one-week noise

4. Is anyone blocked waiting for me? (Design decision, market research, spec clarification)

   • Check JIRA/Slack for "waiting on PM"
   • If yes: This gets done before anything else today
   • Action: 5-minute decision or 15-minute conversation

5. Did we miss something? (Customer issue that should've been escalated, a ship that had a bug, something that doesn't add up)

   • If yes: Quick standup with the team to figure out what happened and how to prevent it
   • Action: Blameless postmortem, not blame assignment

What changes your week's plan:

• Anything marked "at risk" that's a blocker for shipping
• Any metric that suggests your core thesis is wrong (e.g., "we thought low-end users don't activate well, but this cohort is activating fine")
• Any customer churn or deal loss that's a pattern, not a one-off
• Anything that lets you pull forward a planned feature (you got done early, or a blocker cleared)

Everything else gets noted but doesn't change what you're shipping.

Phase 3: Communicate Your Week (10 Minutes)

Now your team needs to know what you're focused on. This is your weekly standup context-setter.

Generate a weekly focus statement using this prompt:

Here's my weekly brief and my triage notes:

[Paste the brief]

[Paste: What's at risk, what changed our plan, what's our focus]

Generate a 3-minute standup opener that:
1. Leads with the headline (what happened)
2. Explains how it changes what we're shipping (if it does)
3. Calls out any blockers we need to unblock
4. Ends with "here's what we're focused on this week and why"

Make it conversational (not corporate), honest (if we had a miss, say it), and forward-looking (less dwelling on problems, more on solutions).

Audience: Product + Engineering team

Paste that into your Monday standup or weekly update message. Takes 2 minutes to run. Your team now knows what happened and why you're focused on X this week.

Connecting the Tools

For this system to work, you need data flowing automatically. Here's the minimum viable setup:

Option 1: Build It With Zapier / Make (1-2 hours setup)

• Trigger: Every Friday at 4pm

• Actions:

  • Pull last week's commits/PRs from GitHub (or closed tickets from JIRA)
  • Export metrics from your analytics tool (Google Analytics, Amplitude, or custom dashboard)
  • Grab top support tickets from Zendesk or Intercom
  • Pull sales metrics from Salesforce (deals closed, pipeline movement)
  • Compile into a Google Doc or shared spreadsheet

• Result: A document you paste into the AI prompt Monday morning

Option 2: Build It With a Script (1-3 hours, if you have eng help)

A simple Python or Node script that:

• Queries GitHub/JIRA API for shipped code
• Hits your analytics API for metrics
• Pulls support tickets from your help desk
• Formats everything into a markdown doc
• Saves to a shared Slack channel, Google Doc, or email

This is more robust if you have complex data sources.

Option 3: Manual Collection (30 minutes, no automation)

If automation feels like overkill, just ask your team on Friday afternoon to post:

• Engineering: "What shipped this week?" and "What's blocking next week?"
• Sales: "Deals closed? Objections we're hearing?"
• Support: "Top 3 issues?"

Paste their responses into the prompt. Less elegant, but it works.

The Weekly Checklist

Print this and keep it on your desk. Every Monday, work through it.

Before Monday 8am:

• Data sources are updated (engineering, metrics, customer, sales, marketing)
• Last week's brief is handy (for week-over-week comparison)

Monday 8am-8:10 (Gather):

• Run the brief prompt
• Skim the output (you're looking for "oh, I didn't know that" items)
• Read any metrics that moved more than 10% in either direction
• Check for customer escalations or deals lost

Monday 8:10-8:20 (Triage):

• Answer the 5 triage questions
• Flag anything that changes the week's plan
• Mark any "waiting on PM" items
• Identify any team blockers or decisions needed

Monday 8:20-8:30 (Communicate):

• Generate the standup opener
• Post to your team channel / standup doc
• Grab a coffee, you're done with the review

During the week:

• One 15-minute sync if something "at risk" got worse
• Otherwise, just ship and review again next Monday

Common Mistakes to Avoid

Mistake 1: The brief becomes the decision. The brief is input to your decision-making, not the decision itself. You still need judgment. If the brief says "churn is up 2%," that's data. You still need to figure out if it's real or noise, and what to do about it.

Mistake 2: You over-automate and lose signal. Temptation: "I want the AI brief to tell me exactly what to do this week." Resist this. The AI is great at surfacing data. You're good at judgment. Keep those separate.

Mistake 3: You treat weekly review like a report. This should be 30 minutes tops and directly informative. If you're spending an hour on weekly review, something's wrong. Either the data sources are too noisy, or you're over-analyzing.

Mistake 4: You miss patterns because you're focused on week-to-week. Every month (last Monday of the month), run a "4-week trend" prompt:

Here are the last 4 weekly briefs:

[Paste all 4 weeks]

What patterns do you see that won't show up in a single week? Trends, emerging issues, strategic shifts?

What should I be watching differently going forward?

This catches slow-moving problems that look fine in isolation.

Integrating With Your Roadmap

The weekly review feeds your roadmap work. Every 4 weeks, after your monthly trend review, ask:

Based on the last month of weekly reviews:

[Paste monthly trends summary]

These were our planned priorities for Q2: [paste roadmap]

Should we adjust what we're shipping? What's over-indexed, what's under-indexed? What did we learn that changes our strategy?

This keeps your roadmap grounded in what you're actually learning from customers and metrics, not just what looked good in the planning meeting.

Why This Works

This system works because:

1. It's regular. Every Monday, you know exactly what to do and it takes 30 minutes. No "how do I even start?" paralysis.

2. It surfaces surprises. The AI brief catches things you might miss skimming messages. A metric dropped 8% but you didn't notice because it was buried in 20 other messages.

3. It creates constraints. 30 minutes forces you to prioritize what matters. You can't dive into analysis rabbit holes.

4. It's repeatable and improves. The first week feels rough. By week 4, you're getting better at the triage questions. By month 2, you're spotting patterns automatically.

5. It frees up time for what only PMs can do. Customer conversations, strategic thinking, unblocking teams. Not "what happened last week" synthesis.

Download the companion artifact: the exact checklist, the prompts copy-pasted and ready to go, and a template data-gathering form to share with your team every Friday.

Sources: Anthropic Claude, GitHub, Atlassian Jira, Salesforce, Zapier, Make.