Guide
How to write OKRs with ChatGPT
ChatGPT or Claude can draft quarterly OKRs that hold up, if you give it the quarter's priorities and the current numbers and ask for a strict shape. One to three objectives, each a qualitative endpoint someone could repeat from memory. Two to four key results per objective, each a metric moving from a baseline to a target by a date. Anything that's really a task ("launch the new onboarding flow") moves to an initiatives list under the objective. Each key result marked committed or aspirational, with one owner, a person, not a team. And a counter-metric next to any key result that could be gamed, such as trial sign-ups against trial-to-paid. Then the team checks the baselines are real before anyone commits to them.
Short answer
- One to three objectives.
- Key results: a metric from a baseline to a target.
- Tasks go to initiatives.
- One owner each, and a counter-metric where needed.
The brief to copy
OKR brief
"Draft next quarter's OKRs.
One to three objectives, each a qualitative endpoint a team member could repeat from memory.
Two to four key results per objective, each '[metric] from [baseline] to [target]'. Use only baselines I give; otherwise write [baseline: check].
Test each key result: if we did this and nothing else changed, would we celebrate? If not, move it to Initiatives.
Mark each key result committed or aspirational, give it one owner (a person), and add a counter-metric if it could be gamed.
Priorities and current numbers: [paste]"
A worked example
Illustration: a made-up business. A ten-person software startup's priority for the quarter is getting new customers set up faster. Day-1 onboarding completion is 34% today. The draft, in part:
From the OKRs
Objective: make the first day something new customers finish.
KR 1 (committed, owner: Mia): day-1 onboarding completion from 34% to 60%.
KR 2 (aspirational, owner: Tom): median time to first booking from [baseline: check] to under 10 minutes.
KR 3 (committed, owner: Ana): support tickets about setup from [baseline: check] to half that. Counter-metric: customer satisfaction on setup tickets doesn't fall.
Initiatives (moved from the first draft): launch the new onboarding checklist; record three setup videos.
The founder pulls the two missing baselines from the product analytics before the planning meeting, so no target rests on a guess.
Before you commit
- Fill every [baseline: check] from real data.
- Move every task to initiatives.
- One person owns each key result; they report it, not do all of it.
- Add a counter-metric wherever a number can be pushed the wrong way.
- Save the brief for next quarter.
Next step: a project that knows the plan
OKRs are easier to keep when your AI already knows them. A Claude project for founders keeps your plan, metrics and rules in one place, so updates and decisions start from them.
Skills for this
- ManagementQuarterly OKRs from a strategyDrafts quarterly OKRs: 1 to 3 objectives, measurable key results with baseline, target and owner, and a check that no KR is a task.
- ManagementQuarterly planning packBuilds a quarterly plan: last quarter's review, real capacity, a ranked priority list with a cut line, dependencies, risks, and commitments vs stretch.
- FinanceMonthly KPI summary for leadershipWrites a monthly KPI summary: a headline, each KPI vs last month and target with the reason, and 3 things to act on. Never invents numbers.
- ManagementDecision memoWrites a short narrative decision memo: recommendation first, context, options with trade-offs, risks, and the ask with a decider and deadline.
Questions
Can ChatGPT write OKRs?
It can draft them, if you give it the quarter's priorities and current numbers and ask for the shape: one to three objectives, two to four key results each written as a metric from a baseline to a target, outputs moved to an initiatives list, one owner per key result, and a counter-metric for any key result that could be gamed.
What's the difference between a key result and a task?
A key result is a change you'd celebrate on its own: "onboarding completion from 34% to 60%". A task is something you do: "launch the new onboarding flow". Ask of each key result, "if we did this and nothing else changed, would we celebrate?" If not, it's an initiative.
How many OKRs should a small team have?
Few. Google's re:Work guide says to pick just three to five objectives, with about three key results each, and warns that more can over-extend teams and diffuse effort. For a small startup team, we suggest one to three objectives: fewer is easier to keep in everyone's head.
AI skill
Have your AI draft next quarter's OKRs from your strategy
Put this skill in AI Skill Pod and everyone on your team gets it in their Claude, Cursor or ChatGPT (Pro, Business, Enterprise): one invite link, free to start. Quarterly OKRs from a strategy has your AI draft quarterly OKRs that hold up: one to three objectives, key results from a baseline to a target, outputs moved to initiatives, committed or aspirational, one owner each, and a counter-metric where a result can be gamed.
14 days free · No card needed
Sources
- Google re:Work: Set goals with OKRs — "Pick just three to five objectives - more can lead to over-extended teams and a diffusion of effort"; about three key results each; measurable; the sweet spot for an OKR grade is 60 to 70% (read 2026-11-05)
- What Matters: OKR meaning, definition and example — committed and aspirational OKRs (read 2026-11-04)