QBR Chain: Customer Value → Presentation → Expansion Decision & Follow-Up

What this chain helps you accomplish

Build a customer-centered quarterly business review (QBR) that demonstrates supported progress toward the customer’s goals, confronts problems fairly, creates a useful discussion, and records real next steps. Explore expansion only when customer needs and account health justify it.

Six connected AI prompts produce a QBR Evidence & Customer Health Brief, a QBR Agenda & Value Narrative, an Expansion Readiness & Discovery Plan, a QBR Presentation Script & Speaker Notes, a Post-QBR Evidence & Commitments Record, and a QBR Follow-Up & Account Action Plan. Each step consumes the latest seller-approved outputs of previous steps.

The output is a meeting-ready, source-checked outline and script—not a fabricated slide deck or automatically delivered presentation. Health metrics, adoption and realized financial value are treated as different measures; unsupported ROI or customer outcomes stay UNKNOWN. The script adapts to executives, operational stakeholders, and accounts where service recovery should take priority.

Promptifi supplies reusable instructions for the seller’s own AI tool. It does not connect to customer systems, retrieve live health data, attend the QBR, send emails, create calendar invitations, or carry out proposed account actions.

When to use it

• Before a scheduled QBR or executive business review: You have a customer objective, an accountable seller/CSM, and available performance, usage or service information that requires careful synthesis.
• For adoption and value conversations: You need to compare original success criteria against observed usage, customer-reported results, unresolved friction and planned improvements.
• Before an expansion discussion: An AE or CSM wants to determine whether an adjacent use case, additional team, capability or license need is actually supported by customer evidence.
• When an account is at risk: Declining engagement, unresolved support issues or a missed milestone makes a credible improvement plan more important than a pitch.
• After the actual QBR: You need an accurate summary, confirmed versus proposed actions, and a follow-up draft based on the meeting—not on what the presentation script predicted.

Do not use it to invent customer health results, automatically describe license utilization as ROI, or assume that upsell is always appropriate. If there is insufficient outcome evidence, the presentation should surface the missing measurements and ask the customer to validate their priorities. If the meeting did not occur, stop before post-QBR analysis.

A QBR is a collaborative business conversation. Its timing, duration, audience and depth should fit the customer; an artificial fixed two-hour follow-up rule is not required.

Before you begin

Minimum AI tool: Any organization-approved conversational AI tool that accepts text, such as ChatGPT, Claude or Gemini. A premium plan, CRM connection, customer-success platform, web browser, PowerPoint integration, or transcription tool is not required. If the tool cannot actually access customer files, manually paste authorized factual excerpts; a file name or URL is not evidence the AI read it.

Gather these inputs. Replace each double-brace placeholder in the Starting QBR Context with authorized facts, or state NONE / UNKNOWN when information is absent. Do not guess.

• {{CUSTOMER_ACCOUNT}} — Required. Actual customer/contract name and business context, from approved account records. Example: Bayfield Fleet Services, regional transportation operator. If account identity is unclear, stop.
• {{REVIEW_PERIOD_AND_DATE}} — Required. Period being reviewed, prior comparison period if available, and actual meeting date/time from the invite. Example: July–September 2026 QBR; Nov. 6 meeting. Do not compare unlike periods without explaining it.
• {{BUYER_ATTENDEES_AND_ROLES}} — Required as a field; may contain UNKNOWN. Verified invited buyer participants, their job functions, stated priorities and executive/operational mix. Example: Operations VP and Fleet Systems Manager; decision authority unconfirmed. Source: verified meeting invite or notes; do not invent executives.
• {{SELLER_TEAM_AND_SENDER}} — Required. Actual AE/CSM owners and intended follow-up sender, from internal team assignments. Example: account executive and customer success manager, names withheld for draft. If no sender known, label emails DRAFT ONLY.
• {{SOLUTION_SCOPE_AND_LIMITS}} — Required. Current licensed product/service, modules or services in use, actual delivery obligations and approved capability limits. Source: contract/approved product material. Example: service request visibility dashboard, no verified ERP integration.
• {{ORIGINAL_CUSTOMER_GOALS}} — Required for claims about success; may be unverified. Specific onboarding/business goals, success plan or buyer-stated measures, each with source ID E#. Example: E1 — customer success plan, June 2026 — improve visibility into escalated service requests; no quantitative baseline. If missing, say GOALS UNCONFIRMED.
• {{PERFORMANCE_AND_VALUE_EVIDENCE}} — Optional; necessary to quantify business impact. Dated outcomes, baseline, current measure, unit, period, denominator, source and methodology, including customer validation where present. Example: E2 — Q3 report — 58 active users out of 100 licensed; not a cash-savings measure. If absent, enter NONE; no ROI estimate or impact figure may be invented.
• {{ADOPTION_AND_USAGE_DATA}} — Optional. Seats, active users, feature use, comparisons, period and precise metric definition, with E# source/date. Example: 58 monthly-active named users/100 assigned seats for September. If unknown, do not infer health from usage alone.
• {{HEALTH_AND_STAKEHOLDER_EVIDENCE}} — Optional. Defined customer health-score method, renewal risk signals, stakeholder feedback, engagement, survey comments or sponsor changes. Example: E3 — CSM note — sponsor requested training refresh; whether a red/yellow/green score is valid is unknown. If there is no score methodology, don't invent one.
• {{OPEN_ISSUES_AND_REMEDIATION}} — Optional but important for expansion decision. Support tickets, security/service problems, adoption barriers, owners, verified status, target responses and commitments. Example: E4 — support record — issue S-12 remains open, engineering investigation ongoing; date uncommitted. If no record exists, say UNKNOWN rather than 'no issues'.
• {{PRIOR_QBR_ACTIONS}} — Optional. Documented prior commitments, owner and agreed due date, actual completion evidence. Example: E5 — prior QBR — seller agreed to deliver training notes, status completed per ticket. If missing, NONE.
• {{COMMERCIAL_AND_EXPANSION_CONTEXT}} — Optional. Actual renewal date, current contract/license scope, buyer-raised adjacent needs, permission to discuss expansion, approved offerings and commercial restrictions. Example: hypothetical second-team use case under consideration; no budget, sponsor or buyer request. If absent, expansion must remain a hypothesis.
• {{APPROVED_CUSTOMER_PROOF}} — Optional. Authorized quotes, case-study statements or product proof, with sharing permissions and exclusions. Example: E6 — customer-approved statement about user training; no financial endorsement. If NONE, do not invent proof.
• {{MEETING_GOALS_AND_FORMAT}} — Optional. Actual planned duration, presentation style, QBR desired outcome, allowed materials and customer preferences. Example: 45-minute discussion with 15 minutes for questions. If not stated, the AI may propose a flexible agenda, not claim the buyer approved it.
• {{DATA_PERMISSIONS_AND_SAFETY}} — Required. Company AI-use rules, whether customer information can be processed, confidentiality constraints, and recording/transcription policies. Example: approved enterprise AI only; redact customer identifiers. If unknown, do not paste restricted records.
• {{POST_QBR_NOTES}} — Required only for Prompt 5, after the real QBR. Authorized notes/transcript from the completed meeting labeled Q1, Q2 etc., with speaker/date, direct quotation or paraphrase, decision, objections, exact accepted actions and uncertainty. Example: Q1 — customer Operations VP — Nov. 6 — asked for incident plan before any license expansion. If the meeting has not happened, write NOT YET AVAILABLE and stop after Prompt 4.

Evidence convention: E1, E2, etc. = authorized pre-QBR account record (include document/speaker, date, definition and limitations). Q1, Q2, etc. = real post-QBR statement/note. Distinguish VERIFIED RECORD, CUSTOMER-REPORTED, SELLER OBSERVATION, ASSUMPTION and UNKNOWN. Do not manufacture source IDs with invented contents.

Measurement and ROI rule: Adoption counts, feature usage, satisfaction, health scores and financial value are not interchangeable. Show denominator, observation window and source for metrics. When a financial benefit is claimed, require a comparable baseline, confirmed benefit assumptions and costs. If inputs are missing, state ROI NOT CALCULABLE FROM PROVIDED EVIDENCE. Never treat potential capacity release as realized cash savings without confirmation.

Data protection: Redact customer data where required; use only permitted AI applications. Respect consent and local law for meeting recording and transcription. Avoid putting confidential contracts, customer PII, security issues or legal advice into an unapproved AI service.

How to run the chain

1. Open an approved AI conversation. Fill and paste the Starting QBR Context from Chain Steps, with source-labeled E# evidence and clearly stated unknowns. Keep it in the same conversation through the pre-QBR prompts.
2. Run Prompt 1 to audit goals, observed outcomes, health signals and unresolved issues. Correct missing denominators, false ROI, attribution mistakes and stale data. Save and explicitly approve the QBR Evidence & Customer Health Brief.
3. Run Prompt 2 using approved Prompt 1, to design the QBR agenda, outcome narrative, discussion questions and slide/section blueprint. Confirm the audience and honest gap disclosure; approve the QBR Agenda & Value Narrative.
4. Run Prompt 3 to assess whether expansion is appropriate. Its outcome may be READY TO EXPLORE, DISCOVERY ONLY, DEFER FOR CUSTOMER SUCCESS, or INSUFFICIENT EVIDENCE. It must not force an upsell. Approve the Expansion Readiness & Discovery Plan.
5. Run Prompt 4 to write a full, adaptable QBR presentation script and speaker notes, including how to discuss missed goals and when to omit an expansion pitch. Read it aloud, validate all claims and obtain internal review if necessary. Approve the QBR Presentation Script & Speaker Notes.
6. Pause for the real QBR. Hold the meeting separately. Capture permitted Q# notes with speaker attribution and distinguish what was proposed from what the customer actually accepted. Do not simulate the conversation.
7. Run Prompt 5 after the meeting with the actual Q# notes to verify outcomes, feedback and real commitments. Approve the Post-QBR Evidence & Commitments Record.
8. Run Prompt 6 with all five approved artifacts to write the follow-up and internal action register. Review actual recipients, claims, agreed versus proposed next steps, customer feedback, product obligations and timing. Send or update CRM only through separately authorized tools.

Seller approval syntax: APPROVED — [Exact Artifact Name] — v1. Use only this version in subsequent prompts. When a draft is corrected, review and approve the complete new version. An internal approval does not imply the customer has validated a result, accepted an expansion request or agreed to a milestone.

One chat is preferred. If the context becomes too long or you need another AI conversation, paste the filled Starting QBR Context, full approved prerequisite artifacts and versions, actual Q# notes, and known unknowns using the optional handoff template.

Failure handling: If reliable outcome data are missing, prepare an honest qualitative review with measurement questions. If unresolved issues dominate, prioritize remediation. If no real QBR occurred, stop after Prompt 4; do not fabricate post-QBR actions. If nothing was mutually agreed, the follow-up should invite validation instead of saying 'as agreed'.

Completion standard: A reviewable, evidence-safe QBR narrative/script, an appropriate expansion-or-defer decision, and after the actual meeting an accurate customer email and internal action tracker. The chain does not create a slide file, schedule a meeting, send an email, or alter a customer record.