Strategy & Planning

How to track okr progress weekly with AI

Strategy & Planning3 AI tools7 steps6 friction points

Want an agent to handle this on your live data? See the Starch version →

Tracking OKR progress weekly means collecting status updates from every team or function, assessing whether key results are on track, amber, or off track, and producing something actionable — not just a color-coded slide that gets filed away. For most operators, this lands squarely on their plate because no one else owns the rhythm, and skipping it for even two weeks means the company drifts without anyone noticing until the quarter is already lost.

The workflow feels like an AI problem because the core task is synthesis: you have scattered updates from Slack, Notion docs, meeting notes, and people's heads, and you need to turn that mess into a structured, honest read on momentum. AI is genuinely good at pattern-matching across messy text, drafting summaries, flagging where language is vague ('we're making progress' tells you nothing), and generating the questions you should be asking. That part maps well to what LLMs do.

ChatGPT, Claude, and Gemini can all contribute meaningfully here today — if you're willing to do the data-gathering yourself. You paste in raw updates, the model structures them against your OKR framework, identifies lagging key results, and drafts a summary or talking-points doc. Claude handles longer context well, which matters when you're feeding in a week's worth of team updates. The output quality is genuinely good — the constraint is everything upstream of the prompt.

Skip the manual work → See the Starch version

Strategy & Planning3 AI tools7 steps6 friction points

AI walkthrough

How to do it with AI today

A practical walkthrough using ChatGPT, Claude, and other off-the-shelf LLMs — what they're good at, what you'll have to do by hand.

Tools that work for this

ClaudeChatGPTGemini

Step-by-step

1 Before each weekly review session, manually collect status updates: copy text from your Notion OKR doc, paste Slack threads where teams discussed progress, and pull any relevant numbers from spreadsheets or dashboards into a single document.

2 Open Claude (preferred for longer context) and paste your OKR structure — objectives, key results with their targets, and current actuals — as the first block of your prompt. Include the raw team updates directly below it.

3 Prompt the model to score each key result as on track, at risk, or off track based on the data you've provided, then write a 2-3 sentence assessment for each explaining why. Ask it to flag any key result where the update is too vague to score confidently.

4 Ask a follow-up prompt requesting the top 2-3 blockers the AI inferred from the updates, and what decisions the leadership team needs to make this week to keep at-risk OKRs from slipping further.

5 Copy the structured output into your meeting notes doc or OKR tracking sheet. Run your weekly review meeting using the AI-generated summary as the starting frame, then manually update the doc with decisions and action items from the conversation.

6 After the meeting, paste your own notes back into Claude and ask it to extract action items with owners and due dates, then format them as a bulleted list you can send to the team via Slack or email.

7 At the end of each month, repeat the full process to compile the weekly summaries into a month-in-review narrative — again, manually copying from wherever you stored the prior weeks' outputs.

Prompts you can copy

Here are my Q2 OKRs with targets and this week's team updates. Score each key result as on track, at risk, or off track, and write 2-3 sentences explaining each score. Flag any KR where the update is too vague to score.

Based on these OKR updates, identify the top 3 blockers that could cause us to miss the quarter. For each blocker, name the key result it threatens and suggest one concrete action the team should take this week.

Here are meeting notes from today's OKR review. Extract all action items mentioned, assign each to the person named in the notes, and add the due date if one was stated. Format as a bulleted list.

Compare this week's OKR status to last week's update, which I'm pasting below. Identify any key results that moved from on track to at risk, and summarize what changed.

Turn these four weekly OKR summaries into a single month-in-review narrative, around 300 words, suitable for sending to the leadership team. Highlight what we accomplished, what slipped, and what carries into next month.

Reality check

Where this gets hard

The walkthrough above works — until your numbers change, the LLM hallucinates, or you have to re-paste everything next month.

No live connection to your data — every session starts with manual copy-paste from Notion, Slack, and spreadsheets, which takes 20-40 minutes before you've generated a single insight.

Nothing persists between sessions; the carefully structured OKR template you prompted last week isn't automatically applied this week, so output format drifts and comparisons become unreliable.

The model has no memory of prior weeks, so trend analysis — 'this KR has been amber for three weeks' — requires you to manually re-paste historical summaries every single time.

Action items extracted by the AI exist only in the chat window; there's no automatic way to push them into a task manager or notify the responsible person without another manual step.

If your OKR doc lives in Notion and your metrics live in a spreadsheet, you're stitching the context together yourself — the model only sees what you paste, so gaps in your copy-paste mean gaps in the analysis.

Producing a polished weekly update or board-ready summary still requires reformatting the AI output by hand, and if you want slides, that's a separate tool and another manual process.

Tired of the friction?

Starch runs the whole workflow on live data — no copy-paste, no hallucinated numbers, no re-prompting next month.

See the Starch version →

Starch alternative

The same workflow on Starch

Starch is an agentic operating system — describe your OKR tracking workflow in plain English, and an agent builds the persistent app that runs it weekly against your live data, without you re-running prompts or copying anything manually.

Connect Notion once as a scheduled-sync provider — Starch syncs your OKR pages and databases on a schedule, so the weekly review app always reads current key result data without you exporting anything.

Use the Knowledge Management starter app to house your OKR definitions, historical summaries, and decisions in one searchable place — so 'what did we decide about this KR two months ago?' has a real answer, not a Slack archaeology project.

Tell Starch: 'Every Monday morning, pull this week's Notion OKR updates, score each key result as on track, at risk, or off track, and post the structured summary to Slack' — the agent builds that automation and runs it on schedule.

Meeting Notes captures your weekly OKR review conversation in real time, extracts action items with owners, and archives the session — so the gap between 'we discussed it' and 'someone owns it with a due date' closes automatically.

Starch connects to 3,000+ apps through its integration catalog; if your key results pull from HubSpot deals, Stripe revenue, or Google Calendar activity, the agent queries those live when it builds your weekly status — not from a spreadsheet you maintain by hand.

Presentation Agent — currently in development, request beta access — will let you describe 'a 5-slide OKR progress update for the board' and get a formatted deck built from your actual weekly data, without Sunday-night slides.

Get closed-beta access →

Toolkit