Available Context & Tools

@_platform-references/org-variables.md @_platform-references/capabilities.md

Slack Coaching Query

Goal

Give the user actionable sales coaching in Slack — either specific objection-handling advice or a quick performance snapshot. Coaching should be concrete, not generic.

Intent Patterns

Objection Handling (AI-powered)

Triggered when objection_type is extracted from the message.

Use Claude Haiku to generate concise, actionable advice with 2-3 specific response frameworks or phrases.

System prompt: "You are a sales coaching expert. Give concise, actionable advice for handling sales objections. Include 2-3 specific response frameworks or phrases. Keep it under 200 words."

Response format:

Section: *Handling: "{objection}"*
Divider
Section: AI-generated advice (formatted with line breaks for readability)
Divider
Context: "Based on general sales best practices. Your team's specific win data will improve this over time."

Fallback (no AI available):

Section: *Handling: "{objection}"*
Section: "I'd give you specific advice, but the AI service is temporarily unavailable. General tips:"
Bullet list: general objection handling principles
- Acknowledge the concern genuinely
- Ask clarifying questions to understand the root cause
- Reframe around value and business outcomes
- Share relevant proof points and case studies
Context: "Try again in a moment for AI-powered advice."

Performance Snapshot

Triggered when message matches: "how am I", "performance", "doing", "stats", "metric"

Fetch pipeline data and build a quick scorecard.

Response format:

Section: "Your Performance Snapshot:"
Fields (2x2):
- Active Deals: count
- Pipeline Value: total value
- Meetings This Week: count of meetings since start of current week
- Weighted Pipeline: probability-weighted value
Divider
Context: "For detailed coaching insights, check your weekly coaching digest.\nAsk me 'which deals are at risk?' or 'show my pipeline' for more detail."

General Coaching (context-aware)

Triggered when no specific objection is mentioned and not a performance query.

Generate pipeline-aware tips based on current deal state:

If deals at risk (score >= 60): "You have N deal(s) at risk — consider focused attention there."
If many deals in Discovery (3+): "N deals in Discovery — focus on qualifying and advancing these."
If pipeline looks healthy: "Your pipeline looks healthy. Keep up the momentum!"

Response format:

Section: "Quick Coaching:"
Bullet list of context-aware tips
Divider
Context: Suggested follow-up questions the user can ask:
- "How should I handle budget objections?"
- "Which deals need attention?"
- "Show my performance stats"

Data Sources

Performance metrics: execute_action("get_pipeline_snapshot", { owner: slack_user_id })
Active deals: execute_action("list_deals", { status: "active", owner: slack_user_id })
Risk scores: execute_action("get_deal_risk_scores", { owner: slack_user_id })
Meetings: execute_action("list_meetings", { owner: slack_user_id, days_back: 7 })

Week Calculation for Meeting Count

const now = new Date();
const weekStart = new Date(now);
weekStart.setDate(now.getDate() - now.getDay()); // Sunday of current week
const thisWeekMeetings = meetings.filter(m => new Date(m.start_time) >= weekStart);

Response Constraints

Objection advice: max 200 words — concise beats comprehensive in Slack
Include 2-3 specific response frameworks, not abstract principles
Performance snapshot: use same currency formatting as pipeline query (£/$ with K/M)
General coaching: max 3 tips — focus on the most impactful actions
Always end with suggested follow-up questions to drive engagement
Tone: direct, encouraging, specific — avoid generic motivational phrases

Error Cases

AI unavailable for objection advice: Show structured fallback with general principles (never return an empty response)
No pipeline data for performance: Show what's available, note missing data explicitly
No deals at all: Skip deal-specific coaching tips, focus on prospecting advice