k2f-business-report

K2F Business Report — AI-Native Aesthetic Design

Purpose

Teach agents to blueprint business reports before layout: a decision-oriented story from evidence to action, honest about data gaps, and hierarchy that lets executives grasp performance in minutes—not generic prose or chart dumps without interpretation.

Built on K2F

This skill produces business reports as .K2F packages. Follow the k2f skill for authoring: unpack or init a workspace, edit the semantic tree, run pack/verify, then export. K2F keeps agents on meaning (section roles, metrics, recommendations)—not pixel coordinates—while the engine compiles layout, typography, and theme into consistent output across PDF, DOCX, PPTX, and other formats.

Before you build, install the core skill: npx skills add suzheng/k2f --skill k2f.

When to Use

• Weekly, monthly, or quarterly performance reviews from spreadsheets, CRM exports, or pasted metrics.
• Diagnostic reports that name weak areas, severity, and improvement strategies from user-supplied data.
• Board- or leadership-ready summaries where trends, risks, and next steps must align with one narrative thread.

Workflow

Step 1. Generate the content_and_design.md

Single blueprint for semantic content and aesthetic intent. Use labeled blocks the k2f skill can map to sections and visuals later.

Sub-steps

1. Report charter — Audience (exec, ops, board), period covered, primary question (“Are we on track?” “Where are we leaking?”), and cadence (weekly pulse vs deep quarterly). List data sources the user actually provided; mark gaps instead of inventing figures.

2. Clarify scope (when inputs are thin) — Which metrics matter, comparison baseline (prior period, target, budget), and whether the reader needs full detail or highlights-only. Do not expand scope beyond the charter.

3. Information architecture — Default spine (adapt or shorten for weekly pulses):

   • Cover / title block: report name, period, prepared for, date.
   • Executive summary: headline verdict, 3–5 bullet takeaways, top risk and opportunity—numbers only where user data supports them.
   • Key metrics: small set of KPIs with current value, delta vs prior period and vs target; one line of interpretation each.
   • Performance trends: time-series or segment story; prefer chart roles over long paragraphs when shape matters (growth, decline, volatility).
   • Findings / weak areas: prioritized by severity; each finding = observation + business impact + plausible driver (labeled inference if not in data).
   • Recommendations: 3–6 initiatives tied to findings; per initiative—objective, 3–5 prioritized actions, expected impact (H/M/L), horizon, success metrics.
   • Appendix intent (optional): tables, definitions, methodology—only if the audience needs audit trail.

4. Analytical discipline — Separate fact (from supplied data), interpretation (your reading), and recommendation (proposed action). Flag anomalies and data quality issues. When severity is assigned, state the rule (e.g., negative growth, missed target, bottom-quartile segment). Never fabricate benchmarks or statistics.

5. Recommendation shape — Each strategy names a clear program, links to a specific weak area, and ends with measurable success criteria. Cut generic advice (“improve marketing”) unless rewritten as concrete actions with owner-type hints when the user named teams.

6. Design intent — Document tone, hierarchy, density, palette for the engine:

   • Tone: confident, neutral-analytic; suitable for skimming in a meeting. No hype, no apology essays.
   • Hierarchy: executive summary and KPI strip highest contrast; findings use severity as visual rank (critical before minor); recommendations as scannable cards or numbered blocks.
   • Density: bullets and short paragraphs; one idea per block; tables for metric grids; reserve narrative depth for summary and “so what” under each finding.
   • Palette: professional corporate (deep neutral text, one accent for deltas and severity); WCAG AA; charts limited to a coherent family (trend line, ranked bars, contribution)—no decorative dashboard clutter.

Rules & constraints

• Decision thread: every section answers “what should leadership do differently?”—cut interesting but non-actionable analysis.
• Severity before volume: fewer, ranked problems beat long undifferentiated lists.
• Chart vs prose: use chart intent when the insight is shape or comparison; use prose when causality or nuance needs words.
• Honesty gate: missing data → NEEDS USER INPUT; do not backfill with plausible numbers.
• Audience fit: weekly reports emphasize delta and actions; quarterly reports add trend context and initiative portfolios.

Common failure modes

• Data dump without verdict → add executive summary with explicit headline and ranked takeaways.
• Recommendations disconnected from findings → map each initiative to a finding ID.
• Vague weakness (“sales could be better”) → replace with metric, period, and threshold.
• Equal weight to every KPI → designate 4–7 primary metrics; demote the rest to appendix.
• Essay trends where a line chart tells the story → switch block to chart role + one interpretive sentence.
• Fabricated growth rates or benchmarks → remove or label as assumption pending user confirmation.
• Generic strategy boilerplate → rewrite as named program with actions and metrics.

Quality checklist

• Charter states audience, period, and primary question.
• Executive summary readable alone; takeaways match body findings.
• Each KPI has value, comparison, and one-line “so what.”
• Findings ordered by severity; impact stated for each.
• Every recommendation links to a finding and includes success metrics.
• Facts traceable to user inputs; gaps explicitly marked.
• Design intent specifies hierarchy for summary, KPIs, findings, and recommendations.

Step 2. Use the content_and_design.md and k2f skill to generate a K2F file and export to an appropriate file format.

Follow k2f to instantiate sections and chart roles from the blueprint, apply design intent, verify, and export the executive-ready deliverable.