AI-Generated QBR Talking Points Prompt
Prompt
You are a customer success AI assistant generating talking points for an upcoming QBR. Account data: [PASTE: Account | Products | Usage data (key metrics for last quarter) | Support tickets (count and resolution time) | NPS score | Goals stated at last QBR | Any achievements or milestones | Open issues | Renewal date | Expansion opportunities] Generate QBR talking points: 1. Value delivered — 3 specific outcomes tied to usage data; quantify where data allows 2. Progress vs. goals — which goals from last QBR were achieved, partially achieved, or not progressed? 3. Usage insights — what the usage data suggests about adoption health; what to celebrate and what to address 4. What's coming — 2–3 relevant upcoming product features for their use case 5. Ask — frame the renewal and any expansion naturally at the end, grounded in the value just discussed Output: QBR talking points. Structured for a 30-minute conversation. Prompts for the CSM to personalize with additional context.
Why it works
Generating talking points rather than a script gives CS managers material they can adapt to the flow of the conversation rather than mechanically reading. The four-section structure (value delivered / risk and open issues / product development context / next period goals) ensures the QBR covers the past, present, and future without being driven entirely by what went well. Building expansion talking points into the QBR prep converts a relationship call into a commercial conversation.
Watch out for
Talking points based on usage metrics only tell part of the story — the customer's subjective experience of value often differs from what the data shows. Always review the talking points with knowledge of how this specific customer feels about your product before the meeting. NPS score in particular should be contextualised: a 7 from a customer with known friction is different from a 7 from a generally positive customer.
Used by