QBR Generator
Transforms raw account data exports into a quarterly business review package: health narrative, spend trend analysis, adoption story, risk flags, and a slide-by-slide deck outline.
Inputs
orders.csv: order_id, date, category, amount, fulfillment_statussupport.csv: ticket_id, opened, closed, severity, categoryadoption.csv (optional): feature, monthly_active_users, month- Account context: contract value, renewal date, key stakeholders, prior-quarter commitments
Sample synthetic inputs live in references/sample_data/.
Workflow
- Run
scripts/aggregate.py to compute quarter-over-quarter spend, category mix shift, order velocity, ticket volume and mean-time-to-resolution.
- Score account health using the rubric in
references/health_rubric.md (spend trend 40%, adoption 30%, support burden 20%, engagement 10%).
- Write the narrative in this order: headline metric, what changed this quarter, why it changed, risk or expansion signal, recommended next commitment.
- Produce a deck outline: title slide, executive summary, spend and adoption trends, support review, roadmap alignment, proposed next-quarter commitments.
- Flag every claim in the narrative against a computed number. No unverifiable statements in customer-facing output.
Output format
Markdown package with three sections: ## Executive Summary (under 150 words), ## Health Analysis (metrics table plus narrative), ## Deck Outline (one line per slide with the single number that slide must carry).
Guardrails
- Never invent metrics. If an input file is missing a field, state the gap and proceed with available data.
- Attribute every trend to a computed delta, not to intuition.
1---2name: qbr-generator3description: Generate a QBR-ready account health narrative, executive summary, and deck outline from raw account data (orders, spend, support tickets, adoption metrics). Use this skill whenever the user mentions QBRs, quarterly business reviews, account reviews, executive readouts, renewal prep, or asks to turn account data exports into a customer-facing summary, even if they only provide a CSV and say "prep this account."4---56# QBR Generator78Transforms raw account data exports into a quarterly business review package: health narrative, spend trend analysis, adoption story, risk flags, and a slide-by-slide deck outline.910## Inputs1112- `orders.csv`: order_id, date, category, amount, fulfillment_status13- `support.csv`: ticket_id, opened, closed, severity, category14- `adoption.csv` (optional): feature, monthly_active_users, month15- Account context: contract value, renewal date, key stakeholders, prior-quarter commitments1617Sample synthetic inputs live in `references/sample_data/`.1819## Workflow20211. Run `scripts/aggregate.py` to compute quarter-over-quarter spend, category mix shift, order velocity, ticket volume and mean-time-to-resolution.222. Score account health using the rubric in `references/health_rubric.md` (spend trend 40%, adoption 30%, support burden 20%, engagement 10%).233. Write the narrative in this order: headline metric, what changed this quarter, why it changed, risk or expansion signal, recommended next commitment.244. Produce a deck outline: title slide, executive summary, spend and adoption trends, support review, roadmap alignment, proposed next-quarter commitments.255. Flag every claim in the narrative against a computed number. No unverifiable statements in customer-facing output.2627## Output format2829Markdown package with three sections: `## Executive Summary` (under 150 words), `## Health Analysis` (metrics table plus narrative), `## Deck Outline` (one line per slide with the single number that slide must carry).3031## Guardrails3233- Never invent metrics. If an input file is missing a field, state the gap and proceed with available data.34- Attribute every trend to a computed delta, not to intuition.
Prashanthvara