Audience-Adaptive Narrative
When to Activate
Trigger when:
- User mentions "board memo", "board briefing", "board presentation"
- User mentions "client letter", "client report", "investor update"
- User mentions "PM deep-dive", "portfolio manager analysis", "detailed factor analysis"
- User specifies an audience for a report or analysis
Audience Tiers
Tier 1: CIO / Board / Executive
Trigger words: "board", "CIO", "executive", "C-suite", "stakeholder", "senior leadership"
Format:
- Executive summary: 3-5 bullet points maximum
- Lead with conclusion and recommendation
- One summary table only (no raw data tables)
- No jargon — plain business English
- Charts: 1-2 high-level (pie chart for allocation, line for performance)
- Length: 1-2 pages equivalent
Tone: Authoritative, decisive, action-oriented
Tier 2: Portfolio Manager / Analyst
Trigger words: "PM", "portfolio manager", "analyst", "deep-dive", "detailed", "factor-level"
Format:
- Full detail with supporting evidence
- Multiple data tables (factor returns, IC scores, risk metrics)
- Statistical measures (t-stats, p-values, confidence intervals)
- Technical terminology acceptable
- Charts: Multiple (scatter plots, heatmaps, time series)
- Length: 3-10 pages equivalent
Tone: Technical, evidence-based, nuanced
Tier 3: Client / Investor
Trigger words: "client", "investor", "external", "prospect", "RFP"
Format:
- Context-heavy with educational framing
- Explain what metrics mean, not just what they are
- Avoid acronyms or define them on first use
- Comparative benchmarks always included
- Charts: Clean, labelled, with annotations explaining significance
- Length: 2-4 pages equivalent
- Include regulatory disclosures where applicable
Tone: Professional, reassuring, transparent
Workflow
- Detect audience from user request using trigger words above
- If no explicit audience, apply default: Tier 2 for analytical queries, Tier 3 for reports
- Apply the corresponding format, tone, and length constraints to the response
- If generating for multiple audiences (e.g., "board summary and detailed appendix"), produce sections at different tiers
Default Behaviour
If no audience is specified, default to Tier 2 (PM/Analyst) for analytical queries and Tier 3 (Client) for report generation queries.
Stopping Points
- This is a modifier skill applied during response generation. No standalone stopping points.
- Inherit stopping points from the calling workflow.
Output
Response formatted according to the detected audience tier's format, tone, and length constraints.
1---2name: audience-adaptive-narrative3description: Use this skill to adapt response depth, formatting, and language to the target audience. Activate when the user specifies an audience like "board memo", "client letter", "PM deep-dive", or when generating reports for different stakeholders, even if they don't explicitly name an audience tier.4---56# Audience-Adaptive Narrative78## When to Activate910Trigger when:11- User mentions "board memo", "board briefing", "board presentation"12- User mentions "client letter", "client report", "investor update"13- User mentions "PM deep-dive", "portfolio manager analysis", "detailed factor analysis"14- User specifies an audience for a report or analysis1516## Audience Tiers1718### Tier 1: CIO / Board / Executive1920**Trigger words**: "board", "CIO", "executive", "C-suite", "stakeholder", "senior leadership"2122**Format**:23- Executive summary: 3-5 bullet points maximum24- Lead with conclusion and recommendation25- One summary table only (no raw data tables)26- No jargon — plain business English27- Charts: 1-2 high-level (pie chart for allocation, line for performance)28- Length: 1-2 pages equivalent2930**Tone**: Authoritative, decisive, action-oriented3132### Tier 2: Portfolio Manager / Analyst3334**Trigger words**: "PM", "portfolio manager", "analyst", "deep-dive", "detailed", "factor-level"3536**Format**:37- Full detail with supporting evidence38- Multiple data tables (factor returns, IC scores, risk metrics)39- Statistical measures (t-stats, p-values, confidence intervals)40- Technical terminology acceptable41- Charts: Multiple (scatter plots, heatmaps, time series)42- Length: 3-10 pages equivalent4344**Tone**: Technical, evidence-based, nuanced4546### Tier 3: Client / Investor4748**Trigger words**: "client", "investor", "external", "prospect", "RFP"4950**Format**:51- Context-heavy with educational framing52- Explain what metrics mean, not just what they are53- Avoid acronyms or define them on first use54- Comparative benchmarks always included55- Charts: Clean, labelled, with annotations explaining significance56- Length: 2-4 pages equivalent57- Include regulatory disclosures where applicable5859**Tone**: Professional, reassuring, transparent6061## Workflow62631. Detect audience from user request using trigger words above642. If no explicit audience, apply default: Tier 2 for analytical queries, Tier 3 for reports653. Apply the corresponding format, tone, and length constraints to the response664. If generating for multiple audiences (e.g., "board summary and detailed appendix"), produce sections at different tiers6768## Default Behaviour6970If no audience is specified, default to Tier 2 (PM/Analyst) for analytical queries and Tier 3 (Client) for report generation queries.7172## Stopping Points7374- This is a modifier skill applied during response generation. No standalone stopping points.75- Inherit stopping points from the calling workflow.7677## Output7879Response formatted according to the detected audience tier's format, tone, and length constraints.
snowflake-labs