How is this different from just making a list of what to build next quarter?

A list ranks features; this produces a plan with a frame, success metrics, and — most importantly — an explicit cut list of what you're deliberately not doing and why. It also traces every commitment back to your strategy doc, your market read, and synthesized customer evidence, so the plan survives scrutiny instead of being a gut-feel wishlist that shifts with whoever argues hardest.

Does Claude decide the priorities for me?

No. Claude synthesizes your inputs, generates candidates, and lays them out in a scored table with its reasoning — but it deliberately does not hand down a verdict. You own every commit and every cut. The scoring organizes the decision; the call is yours to make and defend to your team, which is the whole point of the exercise.

Why is the cut list treated as the most important part?

Because saying what you won't do is what makes a plan real — a plan that commits to everything has prioritized nothing. The cut list, with a one-line reason attached to each item, is what stops the quarter from sprawling and what you point to when a stakeholder asks why their feature isn't in. It's the hardest artifact to write and the most valuable one to keep.

Can I trust Claude's effort and capacity estimates?

Treat them as a starting shape, not a final number. Claude's effort estimates are educated guesses with no real view into your codebase's friction or your team's velocity, so verify them with whoever actually builds the work before committing the quarter. The estimates are useful for seeing whether the plan roughly fits your capacity, not for promising a delivery date.

Run quarterly planning and prioritization end to end

when to reach for this

Quarterly planning is where strategy meets reality, and most teams botch the meeting it happens in. The plan ends up shaped by whoever argued hardest, or it's a gut-feel wishlist with no thread back to the strategy — twelve initiatives, no priorities, and nobody able to say what you're deliberately not doing. This is the orchestration playbook that pulls every foundation together: it feeds Claude your strategy doc, your competitive read, and your synthesized customer feedback, then runs a disciplined planning pass that ends in a one-page plan you can defend line by line. It runs entirely in Claude Desktop — you drag your docs into the file pane, talk through the prioritization in the chat, and review every draft as an accept/reject diff. Claude does the synthesis and the scoring; the call — especially the cut list — stays yours.

gather this first

Your product-context.md (from the idea-to-spec / strategy work) — the north star, who you serve, and what you optimize for this year. This is the frame the whole plan must trace back to.
Your market-landscape.md (from the competitor-intel read) — where the market is moving and where rivals are pushing, so the plan is shaped by the outside, not just the inside.
Your synthesized customer feedback (from the feedback-synthesis playbook) — the de-identified pattern of what customers are actually asking for, with distinct-customer counts, not one loud account.
A quarterly-plan.md working doc and an honest read on capacity — how many real engineering- and design-weeks you have this quarter, accounting for support, bugs, and leave.

the workflow

Set the frame before you plan a single thing

Open the folder with your strategy and market docs in Claude Desktop and ask in the chat — no terminal needed. Approve the reads when the 'Ask permissions' prompt appears. Before generating any candidates, make Claude restate the strategy in its own words and name the ONE thing this quarter must move. Planning on a misread frame is the most expensive mistake here — every downstream initiative inherits it. Catch a wrong frame now, while it costs one prompt.

you ask

Read product-context.md and market-landscape.md. Before you suggest anything to do this quarter, restate back to me: our north star, who we serve, what we're optimizing for this year, and — given the market read — the single most important thing this quarter needs to move. Don't propose initiatives yet.

what you get back A short, sharp read-back — "north star: weekly active teams; this quarter the one thing that must move is activation, because the market read shows two rivals shipping faster onboarding." If that frame is wrong or fuzzy, correct it now; everything below is built on it.

If Claude names the wrong 'one thing,' that's the bug-catch the whole afternoon depends on. A plan aimed at the wrong target is worse than no plan — fix the frame before you build on it.
Generate candidate initiatives from the evidence, grouped into themes

Now turn the synthesized feedback and your roadmap signals into a candidate list — and crucially, group it into a handful of themes rather than a flat pile of twenty ideas. Themes are how a plan stays legible to the team: people can hold three bets in their head, not twenty tickets.

you ask

Using the synthesized customer feedback and our roadmap notes, generate candidate initiatives for the quarter that could move [the one thing from step 1]. Group them into 3-4 themes. For each candidate: a one-line description, which theme it serves, and the customer evidence or signal behind it. This is a longlist to choose from, not a commitment.

what you get back A grouped longlist — e.g. three themes (Activation, Retention, Expansion) with 3-5 candidate initiatives under each, every one tied to a real signal ("6 distinct customers asked for guided setup") rather than a hunch. You now have a menu shaped by evidence, not by who's loudest in the room.

If a candidate has no evidence behind it — no feedback, no market signal — flag it. A 'wouldn't it be cool if' idea isn't disqualified, but it should be labeled as a bet, not dressed up as demand.
Run a structured prioritization pass — scoring, not a verdict

Score the candidates so the choices are comparable instead of vibes. Have Claude lay them out in a table — impact, effort, confidence (or full RICE: reach, impact, confidence, effort) — with its reasoning shown per cell, not a ranked verdict. The reasoning is the point: a score you can interrogate is useful; a number Claude hands down is theater.

you ask

Score every candidate initiative in a table on Impact (1-5), Effort (1-5, where higher = more work), and Confidence (1-5). For each score, give me the one-line reasoning behind it. Show your effort estimates as ranges and flag any you're unsure about. Don't rank them or pick winners — I want to see the reasoning and make the call myself.

what you get back A scored table where every cell carries a short rationale — "Impact 4: directly hits activation, the quarter's target; Effort 3-4: touches onboarding flow + a new email, unsure on the email infra; Confidence 2: no data on whether guided setup actually lifts activation." The table organizes the decision; it doesn't make it.

Watch the confidence column hardest. A high-impact, low-confidence initiative isn't a 'do it' — it's a 'de-risk it first.' Low confidence is a signal to run a small test, not to commit the quarter.
Draft the plan — and the cut list that gives it teeth

Now assemble the actual one-page plan: the themes, the committed initiatives, and a success metric for each. Then the most important artifact in the whole exercise — an explicit CUT LIST of what you're deliberately not doing this quarter, and why. A plan with no cut list isn't a plan; it's a wishlist. Review the draft as an accept/reject diff in the file pane so you control exactly what lands in quarterly-plan.md.

you ask

Draft a one-page quarterly plan in quarterly-plan.md: the 3-4 themes, the committed initiatives under each (the ones we're actually doing), and for every committed initiative a single success metric we'll judge it by. Then — separately and explicitly — a CUT LIST: the candidates we are NOT doing this quarter, each with a one-line reason (lower confidence, off-theme, deferred to next quarter). Keep it to one page.

what you get back A tight one-pager: themes → committed initiatives, each with a measurable success metric ("guided setup → +10pts activation rate"), and a named cut list ("NOT this quarter: in-app chat — off-theme; SSO — real demand but only 2 accounts, deferred to Q4"). The cut list is what you'll point to in three weeks when someone asks why their pet feature isn't in.

Make Claude justify every cut, not just every commit. The reasons on the cut list are what protect you from re-litigating each one in every standup for the next three months.
Pressure-test against capacity, then surface the risk and the open questions

A plan that ignores capacity is fiction. Feed Claude your honest capacity number and have it reality-check whether the committed work actually fits — then have it name the single biggest risk to the plan and the open questions you still owe an answer. This is the skeptical pass that turns an aspirational list into something you can commit to out loud.

you ask

We have roughly [N] engineering-weeks and [M] design-weeks this quarter, after support and bug-fixing overhead. Pressure-test the committed initiatives against that capacity: does it fit, and if not, what's the leanest thing to cut or shrink? Then tell me the single biggest risk to this plan and the top 3 open questions I still need to answer before I commit to it.

what you get back An honest fit check — "committed work is ~14 eng-weeks against ~11 available; the leanest cut is to phase the expansion theme to next quarter" — plus a named top risk ("the whole plan assumes activation is the bottleneck; if it's actually retention, the quarter is misaimed") and the real open questions. Now the plan is defensible, not just tidy.

Claude's effort and capacity math are estimates stacked on estimates — useful for the shape of the fit, not the final word. Verify the numbers with whoever actually builds the work before you commit the quarter to them.

make it your own

Lighter cadence (monthly or sprint): collapse to three beats — frame → score → plan-with-cut-list — and skip the full RICE scoring for a quick impact/effort pass. The cut list stays non-negotiable even in the short version; it's the part that earns its keep at any cadence.
Arabic / bilingual team: run the whole pass in Arabic if that's the language your leadership actually decides in — ask Claude in Arabic and have it produce quarterly-plan.md in Arabic. Localize the success metrics and themes to how your team talks about them; a plan people argue over in their working language gets honest scrutiny, a translated one gets nodded through. Adapt, don't translate.
Standing process (Power Track): wire the sequence into a /plan-quarter custom command (see the Features tab) so every quarter runs through the same disciplined gauntlet instead of degrading into whoever-shouts-loudest. Custom commands are an opt-in Power-Track step — on Desktop you can keep the dialed-in prompt chain in a note and run it in sequence each quarter.
No clean inputs yet: if you don't have product-context.md, market-landscape.md, or synthesized feedback, run the Stage-1/2 Founders playbooks first — this orchestration playbook is only as strong as the foundations it inherits. Planning on un-synthesized feedback means planning on the loudest account, which is exactly what this is meant to prevent.

watch out for

Claude proposes priorities; you own the call and the cut list. The scoring table organizes the decision — it doesn't make it. A quarterly plan you can't defend to your team in your own words is worthless, no matter how clean the spreadsheet looks. Treat every score as a starting point to interrogate, and own each commit and each cut as your decision, not Claude's.
Verify capacity and effort with whoever actually builds it. Claude's effort estimates are educated guesses with no view into your codebase's real friction, your team's velocity, or the half-finished migration nobody mentioned. Take the scored effort to your eng and design leads and adjust before you commit — a plan built on imaginary capacity ships late and erodes trust.
The cut list is the most valuable artifact — protect it. It's tempting to quietly drop it because saying no is uncomfortable, but the cut list is the entire point: it's what stops the quarter from sprawling and what you point to when a stakeholder asks why their feature isn't in. A plan that commits to everything has prioritized nothing. Keep the reasons attached to each cut so you don't re-litigate them weekly.
Don't let a tidy scoring table launder a bad strategic bet. A clean RICE table feels objective, but the scores are only as good as the frame in step 1 and the evidence in step 2 — garbage in, confident-looking garbage out. If the 'one thing' is wrong, a perfectly-scored plan just gets you to the wrong place efficiently. The numbers organize judgment; they don't replace it. And since the plan summarizes proprietary strategy and de-identified customer patterns, treat it as internal-only — strip account names and keep it out of anything you wouldn't show the whole company.

you'll end up with A defensible one-page quarterly plan — 3-4 themes, the committed initiatives each with a single success metric, and an explicit cut list of what you're deliberately not doing and why — built in an afternoon instead of a week of meetings, and traceable line by line back to your strategy, your market read, and real customer evidence.

Questions people ask

How is this different from just making a list of what to build next quarter?: A list ranks features; this produces a plan with a frame, success metrics, and — most importantly — an explicit cut list of what you're deliberately not doing and why. It also traces every commitment back to your strategy doc, your market read, and synthesized customer evidence, so the plan survives scrutiny instead of being a gut-feel wishlist that shifts with whoever argues hardest.
Does Claude decide the priorities for me?: No. Claude synthesizes your inputs, generates candidates, and lays them out in a scored table with its reasoning — but it deliberately does not hand down a verdict. You own every commit and every cut. The scoring organizes the decision; the call is yours to make and defend to your team, which is the whole point of the exercise.
Why is the cut list treated as the most important part?: Because saying what you won't do is what makes a plan real — a plan that commits to everything has prioritized nothing. The cut list, with a one-line reason attached to each item, is what stops the quarter from sprawling and what you point to when a stakeholder asks why their feature isn't in. It's the hardest artifact to write and the most valuable one to keep.
Can I trust Claude's effort and capacity estimates?: Treat them as a starting shape, not a final number. Claude's effort estimates are educated guesses with no real view into your codebase's friction or your team's velocity, so verify them with whoever actually builds the work before committing the quarter. The estimates are useful for seeing whether the plan roughly fits your capacity, not for promising a delivery date.