Deal Sentiment Monitor

Different output and purpose. SCAL logs ALL Slack deal channel messages to Pipedrive as activities with AI classification (intent category, stakeholder, CRM action). DSM scores the emotional TONE of messages and alerts on negative sentiment shifts. SCAL tracks what happened; DSM detects how people feel about it. Together they provide both activity tracking and emotional intelligence on deal channels.

positive_momentum (8–10) — buying signals, enthusiastic engagement, deal moving forward. neutral_stable (6–7) — normal business communication, no strong emotional signals. cautious_hesitant (4–5) — hedging language, delayed responses, qualification concerns. frustrated_negative (2–3) — complaints, dissatisfaction, competitor mentions, support issues. urgent_escalation (1) — threats to leave, escalation to management, deal-breaking language.

The Syncer updates the Pipedrive fw_deal_sentiment custom field using a weighted running average: (current_score × 0.7) + (new_score × 0.3). This smooths out single-message noise while still responding to sentiment shifts. A single negative message nudges the average down; sustained negativity drives it down further. The 70/30 weighting means recent history matters more than any individual message.

Alerts fire on two conditions: (1) the new sentiment score is in the frustrated_negative or urgent_escalation range, or (2) the sentiment shift exceeds SENTIMENT_SHIFT_THRESHOLD (configurable). When triggered, the Syncer creates a Pipedrive note with the message context, sentiment analysis, and score, plus sends a Slack thread reply on the original message alerting the team. Normal positive/neutral scores update the field silently.

Sentiment classification is a structured classification task that Sonnet 4.6 handles with high accuracy in ITP testing. The Chain-of-Thought prompt (message_summary + tone_signals BEFORE scoring) provides sufficient reasoning structure. Opus would add cost without measurable quality improvement for this task type. SINGLE-MODEL architecture keeps cost at $0.005–$0.011/message — the lowest per-record LLM cost in the ForgeWorkflows catalog.

Uses a configurable DEAL_CHANNEL_PREFIX convention. Channels named with the prefix (e.g., #deal-acme-corp) map to their corresponding Pipedrive deals via prefix matching and deal search. The Extractor strips the prefix, searches Pipedrive for matching deals, and fetches the current fw_deal_sentiment running average for the matched deal.

Complementary products covering different signals. DSM detects sentiment shifts that may PREDICT stalls before they appear in activity metrics — real-time per-message monitoring. DSD diagnoses WHY deals are stalling based on activity patterns and stage duration on a daily schedule. Different trigger (real-time vs scheduled), different signal (emotional tone vs activity absence). Together they provide both predictive sentiment signals and diagnostic stall analysis.

Complementary products covering different scope. DIA provides comprehensive deal analysis with risk scoring across multiple dimensions on demand. DSM monitors real-time message sentiment as a leading indicator of deal health. Different timing (on-demand vs real-time per-message), different scope (full deal intelligence vs sentiment signal). Together they combine deep analysis with continuous sentiment monitoring.

No. All data comes from two sources: Slack Event API (message payload) and Pipedrive API (deal lookup, custom field update, notes). No web_search, no external data sources, no scraping. This makes the pipeline fast and reliable.

All sales are final after download. Review the Blueprint Dependency Matrix and prerequisites before purchase. Questions? Contact support@forgeworkflows.com before buying. Full terms at forgeworkflows.com/legal.