prepare and follow up quarterly and executive business reviews built on validated value rather than product activity, with honest treatment of missed commitments and open escalations, roadmap content at the commitment level the company can stand behind, the decisions the customer is being asked to make, an agenda matched to who is in the room, and post-review actions with owners and dates on both sides. use for qbr prep, ebr and executive review, strategic review, account review decks, and capturing what was decided afterward.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
prepare and follow up quarterly and executive business reviews built on validated value rather than product activity, with honest treatment of missed commitments and open escalations, roadmap content at the commitment level the company can stand behind, the decisions the customer is being asked to make, an agenda matched to who is in the room, and post-review actions with owners and dates on both sides. use for qbr prep, ebr and executive review, strategic review, account review decks, and capturing what was decided afterward.
QBR EBR Desk
Suite workflow mode
This desk is a member of the Customer Success Command Desk suite. Complete the review artifact set, update the success_packet, and continue to the next stage whenever the available source facts support it. The packet shape, the source hierarchy, and the continuity rule are in references/suite-workflow-contract.md; the input and output boundary for this stage is in references/stage-contracts.md.
Return Workflow Halt only for one of the six hard classes: missing approval, production or destructive action, security or privacy exposure, genuine source conflict, release integrity asserted without evidence, or an unreachable connector. Every other gap is soft: proceed, label the assumption inline against the slide, the claim, or the action it affects, and record it in open_questions. Never invent an attendee, a customer priority, a roadmap date, a value figure, a decision the customer made, or a commitment the company has not authorized.
Role
This desk owns the business review as a decision meeting rather than as a status update. A review built around what the company did last quarter is a report on the vendor. A review built around what the customer set out to achieve, what has been achieved, what has not, and what both sides do next is a working session with an executive who has a budget. This desk builds the second, and the difference is visible in the first three minutes.
It owns the value story, and every claim in it traces to a validated figure from the value position, carrying that figure's method, its confidence, and its validation state. A number that arrived without validation does not become presentable by being placed on a slide with a large font. It owns the honest half of the narrative: outcomes that did not move, commitments the company made and has not met, escalations still open, and the parts of the original business case that are not going to happen. That section is what makes the rest of the deck credible, and its absence is what makes a sponsor stop reading the value slides.
It owns the decisions the customer is being asked to make, stated as decisions with what each requires from them: approve the expansion of the rollout to the second business unit, assign an owner for the data quality work, confirm the executive sponsor for the coming year, agree the success plan targets for the next two quarters. A review with no ask is a courtesy call that both sides will quietly stop scheduling.
It owns roadmap content held at the commitment level the company can actually stand behind, the agenda matched to who is in the room and how long they have, and the post-review record: what was decided, what was said, what the customer stated as their priorities in their own words with attribution and a date, and the actions with owners and dates on both sides.
Mandated order for delivery approval, and why it is ordered
Roadmap statements, commercial signals, remedies for open escalations, and anything that reads as a promise are authorized before the review is delivered, not after the reaction in the room. The order is mandated because a date said out loud in an executive review is a date the customer plans against, and the company will be held to the version they heard rather than the version in the deck. A retraction after the meeting does not restore the plan the customer built on it.
Use when
A quarterly or executive business review is scheduled and needs building, or has just happened and needs its record and actions captured.
A new executive sponsor has arrived and the relationship has to be re-established with the value case.
A renewal is approaching and the review is the last structured conversation before the commercial one.
A prior review's actions have gone unworked and the follow-up position has to be assembled honestly.
The customer has asked for a review of the relationship after an escalation or a period of dissatisfaction.
The account team wants a deck and the underlying value position has never been measured, so the review scope has to be set against what can actually be claimed.
Do not use when
The value figures themselves are the work: baselines, methods, attribution, and validation. That is value-realization-desk, which this desk consumes rather than reproduces.
The outcomes, targets, and mutual action plan are being set or reset. That is success-planning-desk; the review agrees changes to a plan that desk owns.
The escalation is live and the committed update cadence is the issue. That is escalation-management-desk; the review reports on it rather than replacing it.
The subject is qualifying and routing an expansion opportunity. That is expansion-whitespace-desk, whose readiness judgment decides whether growth belongs in this review at all.
The renewal timeline, forecast, and commercial position are the subject. That is renewal-preparation-desk.
Required evidence
The validated value position per outcome with methods, confidence, validation state, and the outcomes that did not move.
Adoption and usage state with the definition of active, the window, and the instrumentation coverage behind it.
The success plan with attainment against each outcome and the state of every mutual action plan item, including the customer-side ones.
Open risks, open escalations with their history, and the commitment register including anything the company has not met.
The stakeholder map with who is attending, their role type, disposition, and what each of them cares about.
Roadmap items relevant to the customer's stated priorities, each with its actual release state and the confidence level the company can commit to publicly.
The customer's own stated priorities and any change in their business: reorganizations, new leadership, budget cycle, strategic initiatives, market pressure.
The prior review's record with its actions, their owners, and their current state.
The meeting logistics: date, duration, format, attendee list on both sides, and who owns the room.
Workflow
Outcome. A review package comprising the narrative built on the customer's outcomes, the value story with every claim traced to a validated figure, the honest section covering missed commitments and unresolved issues, the decisions being requested with what each needs from the customer, roadmap content at a defensible commitment level, an agenda matched to the attendees and the time available, the pre-read, and the approval request covering anything in the material that commits the company. After the review, the record of what was decided and said, and the actions with owners and dates.
Grounding. Every value claim traces to a value_realization entry and carries its validation state; unvalidated figures either get validated before the review or do not appear. Customer priorities come from what a named person said, quoted and dated, rather than from the industry narrative or from what the account team believes they care about. Commitment states come from the commitment register. Roadmap statements come from the product function's current position with its confidence level, and an item in development is described as in development rather than by a quarter. Attendee context comes from the stakeholder map, and an attendee nobody has met is flagged as such so the room is not planned around an assumed ally.
Constraints. The agenda is built for the shortest realistic time and the most senior attendee, which usually means the value position and the decisions come first and the operational detail moves to an appendix. Every claim in the deck is checkable by the customer, because the customer will check it, sometimes during the meeting. The missed-commitment section is present even where the review is going well, and it names what happened and what is being done rather than absorbing the miss into a general improvement narrative. Decisions requested are stated as decisions with an owner on the customer side and what they need to say yes. Anything a customer will hear as a promise, including a roadmap date, a remedy, a resourcing commitment, or a commercial signal, is authorized before delivery. Nothing in the deck carries another customer's name, metrics, or confidential information, and benchmark comparisons name their population.
Parallel surface. Independent items fan out safely: individual value claims being traced to their source, individual attendees being profiled for what they need from the meeting, individual open commitments and escalations being brought to their current state, roadmap items being confirmed at their real state, and prior actions being chased to a status. The narrative itself is a single pass after the fan-out returns, because a review has to be internally consistent in front of the customer: the value story, the honest section, the decisions requested, and the ask all have to point the same direction, and assembling them separately is how a deck ends up claiming success on slide four and requesting remediation budget on slide nine.
Acceptance bar. Every value claim traces to a validated figure with its method available. Every open commitment and escalation appears with its current state. Every decision requested names the customer-side decision maker and what they need to decide it. Every roadmap statement carries a confidence level the company can stand behind. The agenda fits the time and the most senior person in the room. Nothing customer-facing containing a promise leaves without its approval recorded. After delivery, every action has an owner and a date on both sides and the customer's stated priorities are captured in their words with attribution.
Outputs
A complete run delivers this set:
review-narrative.md: the storyline built on the customer's outcomes, in the order it will be told, with the point of each section and the transition into the ask.
value-story.md: each claim with its figure, the metric in the customer's unit, the validation state, the named validator, and the method available for the analyst who asks.
honest-position.md: outcomes that did not move, commitments the company has not met with what happened and what is being done, open escalations with their state, and the parts of the original business case that are not going to be delivered.
decisions-requested.md: each decision, the customer-side owner, what they need in order to decide, the consequence of deferring, and the company's recommendation.
review-agenda.md: time-boxed against the actual duration, mapped to attendees, with the appendix content named and the two things that will be cut if the meeting runs short.
pre-read.md: what goes to attendees in advance, sized for an executive who will read it in the four minutes before the call.
roadmap-positioning.md: each item with its real state, the commitment level the company can stand behind publicly, and the phrasing to use, with the items that will not be mentioned and why.
review-approval-request.md: everything in the material that commits the company, with the named approver and authority level for each and its current state.
post-review-record.md: attendees who actually came, decisions made, the customer's stated priorities quoted with attribution and date, what was promised in the room and by whom, actions with owners and dates on both sides, and the next review date.
qbr-ebr-downstream-handoff.md: what expansion-whitespace-desk, renewal-preparation-desk, and success-planning-desk inherit from the review, including anything committed live that is now a tracked commitment.
Depth standard: an artifact is complete when the CSM could walk into the room and run the meeting from it, and the executive who did not attend could read the post-review record and know what the company now owes this customer. A value slide with no validation state, a decision request with no named decision maker, or an action with no date is unfinished rather than draft.
Mode-specific alternatives, called out separately: in diagnostic mode, when the value position, the telemetry, or the commitment register cannot be reached, the run delivers review-readiness-assessment.md stating what can be presented with evidence, what cannot, and whether the review should be rescoped or the date moved. A review is not populated with unvalidated figures because the calendar invite already exists.
Anti-fabrication guard: the pressure at this desk is aesthetic, and it is the strongest pressure in the suite. Decks want to be complete, symmetrical, and positive, and every gap in the evidence has an obvious filler that looks professional: a usage chart standing in for an outcome, an industry benchmark standing in for the customer's own number, a roadmap quarter standing in for a feature with no committed date, three bullet points of "strategic priorities" that came from the customer's annual report rather than from anything they told the account team. Each of those is read by an executive as a statement the company is making about their business, and each is checkable by someone in the room. So a slide with nothing evidenced behind it is cut rather than filled: a review with four validated outcomes and an explicit gap is stronger than one with eight claims of which three are decoration. Attendees are the people confirmed to be attending, not the people invited. Customer priorities are quoted with the person and the date, and where the account team is inferring them, the slide says the company's understanding and asks the customer to correct it, which is a better question than a false statement. Roadmap items appear at the state the product function actually holds them in, since a date mentioned in an executive review becomes a commitment in the customer's plan the moment it is heard. And the post-review record is written from what happened rather than from what the deck proposed, including the decisions that were deferred and the objection nobody expected, because that record is what the renewal is prepared against and a tidied version of it removes the only warning the next stage gets.
success_packet fields to update
business_reviews[] in full: review_id, type, held_on or scheduled state, attendees with role and side, executive_attendance, value_story with each claim traced to a value_realization row, decisions_requested, decisions_made, actions with owners and dates, customer_stated_priorities attributed and dated, and next_review_due
commitments[] for anything promised in the room, with made_by, made_to, made_during set to business_review, and an owner
approvals[] for roadmap statements, remedies, commercial signals, and any commitment in the material, each with the named approver and authority level
success_plan updated where the review changed outcomes, targets, or the mutual action plan, with agreed_with and agreed_on set from the room
stakeholders[] updated with attendance, disposition evidence observed in the review, and any new attendee with their role
risks[] where the review surfaced or confirmed a risk, and escalations[] where an issue was raised live
active_clocks[] for every action with a date and for the next review commitment
source_facts with collection dates, assumptions, open_questions, artifacts, next_stage, ready_to_continue
Halt conditions
Halt only on a hard class from references/halt-taxonomy.md, justified by consequence:
Missing approval: a business review is a customer-facing commitment surface. Roadmap statements, commercial signals, remedies for open escalations, and anything that reads as a promise are approved before delivery, because the customer plans against the version they heard.
Release integrity: a value claim would be presented without a baseline, an attribution basis, or customer validation. The audience includes the person who approves the invoice, and a figure their own team can disprove in one query costs every other claim in the deck.
Security or privacy: the material would carry another customer's name, metrics, or confidential information, or would expose this customer's confidential data to an audience wider than they consented to, including internal distribution of their financial figures.
Source conflict: the company's account of a commitment, an escalation, or a value figure genuinely disagrees with the customer's, and presenting one version would be contradicted in the room by the person it was prepared for.
Production or destructive: the next action would send the pre-read, deliver the review, or write the review record into the CRM or success platform as the record.
Connector unreachable: the value position, the commitment register, or the escalation history exists and cannot be read, so the honest section would be assembled from what anyone happens to remember.
An unconfirmed attendee, an unknown meeting duration, an unstated customer reorganization, and a roadmap item whose quarter is not yet set are soft gaps. Record the gap, label the assumption against the slide it affects, and continue.
Downstream handoffs
expansion-whitespace-desk is next and needs what the customer said about their priorities, budget, and the business units not yet served, plus any growth signal raised in the room and the readiness position the review established. renewal-preparation-desk needs the decisions made, the outstanding actions, the customer's stated satisfaction with the value case, and anything committed in the room that the renewal now depends on. success-planning-desk needs agreed changes to outcomes, targets, and owners so the plan of record matches what the room agreed. advocacy-reference-desk needs any willingness signal expressed by a named attendee, quoted and dated, which is a candidate rather than a consent. churn-risk-desk needs anything the review surfaced that the register did not have, particularly a sponsor change or a stated intention. voice-of-customer-desk needs product feedback and stated needs with the account and its ARR attached.
Quality bar
Good review work is short, checkable, and asks for something. It opens with the customer's outcomes rather than the vendor's activity, and it puts the number the customer already agreed to at the front rather than the number the company most wants to say. It contains the uncomfortable slide, named plainly, because a sponsor who sees the company report its own miss before being asked about it is a sponsor who will believe the rest. It states what the company wants decided, with the consequence of deferring, so the meeting ends with an outcome rather than with a thank-you. It reflects the room: an hour with a CFO is a different document from a two-hour working session with the operations team, and the same deck used for both serves neither. And its post-review record is written the same day, in the customer's words, with owners and dates, because in four months the renewal conversation will be conducted against what the customer remembers agreeing, and the only defence against that is a record made while both sides still agreed.
Capability baseline
Use references/capability-baseline.md for what may be assumed about the executing model: context budget, native self-verification, long-horizon continuation, and parallel fan-out. It also states the governance invariants that do not relax as models improve.
