When to use it
When the quarter is full of wishes but has no focus and nothing measurable — and you or the team spread yourselves thin. The role is an OKR coach. Output: 2-3 Objectives with 3-5 measurable Key Results each, checked against the usual mistakes (KRs that are really tasks, unmeasurable KRs, overload), plus the initiatives underneath them.
The prompt (copy and paste)
You are an OKR coach. Help me set goals for the quarter. Ask first if you don't have enough input.
CONTEXT: "<company/team/solo, what we do>". THE BIG CHALLENGE THIS QUARTER: "<what matters most to move>". INPUTS/NUMBERS: <current metrics, resources>.
Do this:
1. Propose 2-3 FOCUSED Objectives — qualitative, inspiring, answering "where are we going and why" (not a metric, not a task).
2. For each one — 3-5 Key Results: measurable, with a "from → to" number that makes it unambiguous whether we hit it (an outcome, not "ship a feature").
3. Check your own OKRs against the usual mistakes and fix them: KR-is-a-task, unmeasurable KR, too many objectives, objective = activity.
4. Under each Objective — 2-4 initiatives (what we do to move the KRs) — that's the plan, not the OKRs.
5. Flag the overload risk: is this realistic in one quarter given <resources>, and what gets cut first.
Ambitious but achievable (~70% attainment is the benchmark). If an objective is really a task, say so plainly.
Filled-in example
Context: solo founder, SaaS for freelancers. Challenge this quarter: reach stable revenue. Inputs: 50 paying customers, MRR 150k ₽, churn 9%.
What the AI should come back with: O1 "Turn the product into a source of predictable revenue" — KRs: MRR 150k→300k, paying customers 50→110, churn 9%→5%; O2 "Get new customers to value fast" — KRs: share reaching first value 40%→70%, week-one activation up. A pass over the mistakes (replacing the KR "launch onboarding" with the outcome "share of activated users"); initiatives under each; a warning that "2 objectives is the ceiling for a solo founder — push the referral program to next quarter".
Variations
- Cascade. "Break the company goal down into OKRs for 2 teams so they don't duplicate or contradict each other."
- Check-in. Mid-quarter: "here's the KR progress — where's the risk, what to adjust, what we won't make."
- Personal OKRs. For personal development or career goals over a quarter — same rules.
Pro tips
- The number one OKR mistake is KRs that are tasks: "launch the landing page" is done/not done, not a result. A KR has to measure an OUTCOME ("conversion 2%→4%"). Ask the AI to catch and rewrite those.
- Less is more: 2-3 objectives per quarter, otherwise there's no focus and everything sags. If the AI hands you 5 — make it cut down to what matters.
- OKRs are about direction and ambition (~70% attainment is normal), not KPIs with a bonus for hitting 100%. Keep the initiatives (the "how" plan) separate from the KRs (the "what" result), or you'll drown in tasks again.