Commentary Engine · Universal Pattern

When asked to write commentary on numbers, follow this exact structure. Replace placeholders with the specific case.

The Five-Move Structure

1. The headline (one sentence). What is the single most important thing the reader should walk away with?
2. The story (two to three sentences). What is the trend, in plain language? What changed? What does it mean?
3. The drivers (three bullets). What caused the change? Order by magnitude, not by alphabet.
4. The risk or opportunity (one or two sentences). What is the implication going forward? Where is the asymmetry?
5. The action (one sentence). What is the recommendation, or what decision is needed?

Style Rules

• Lead with the insight, not the data. "Revenue declined six percent" beats "Revenue was twenty-six point seven million euros versus twenty-eight point two million."
• Use ranges, not single points, for forward-looking statements. "EBITA likely between three point seven and six point nine million."
• Attribute every variance to a driver. Never leave a number unexplained.
• Quantify the risk. "Bench rising fifty basis points per month" beats "bench rising."
• End on what the reader should do. Commentary that does not lead to a decision is a status update, not commentary.

Voice Calibration

The commentary inherits the brand voice from the [[brand-voice]] skill if available. If not, default to:

• Confident, not arrogant
• Quantitative, not breathless
• Forward-looking, not backward-narrating
• Honest about what is not known

Anti-Patterns

• "The team continued to focus on..." (vague filler)
• "Strong momentum across the business" (no information)
• "We remain optimistic about our outlook" (no quantification)
• "Delve into the dynamics" (AI-tell)
• Single-point forecasts with no range
• Variances without owners or causes

Example Output (Financial)

Q1 revenue landed at twenty-six point seven million, a six percent decline year on year, driven by softness in the engineering segment.

The miss was concentrated in two large accounts where roll-offs were not replaced fast enough. AI delivery grew eighteen percent and absorbed roughly half the gap. Cost actions taken in March will show in Q2.

Drivers:
- Engineering revenue down four point one million on slower replacement
- AI services up nine hundred thousand, exceeding plan
- One-time legal settlement of two hundred thousand, non-recurring

Risk: if engineering does not stabilize by Q3, full-year guidance moves to the lower half of the range.

Recommendation: hold Q2 reinvestment in AI delivery, accelerate engineering pipeline conversion by mid-May, revisit guidance at half-year.

Adapt for non-financial commentary by replacing the financial nouns with operational, HR, or program metrics.