PM Automation Tool

“An AI agent that turns customer feedback from Slack, CRM, and forums into prioritized insights, instantly visualized in Tableau Next.”

Comment

Inspiration

Product teams don’t struggle to imagine features—they struggle to choose them. Valuable signals are buried in Slack threads and support tickets, and the loudest voice often wins. We wanted to fix that by making prioritization data-driven, transparent, and explainable.

The Problem

  • Scattered signal: Feedback lives everywhere and nowhere.
  • Opaque prioritization: Decisions rely on opinions, not evidence.
  • Lost context: Quotes and nuance disappear in spreadsheets.

The Spark

We asked: “What if the top themes picked themselves—backed by evidence—inside the tools teams already use?” No new portals. No manual sorting. Just clear, explainable insights.

What It Does

  1. Ingests feedback from Slack, forums, and CRM into Salesforce (Feedback objects).
  2. Cleans, deduplicates, and clusters using NLP to group related requests.
  3. Scores each cluster by Frequency, Recency, and VIP% for explainable prioritization.
  4. Streams data into Salesforce Data Cloud for governed, analytics-ready insights.
  5. Visualizes in Tableau Next: dashboards showing top clusters, trend lines, and real customer quotes.
  6. Slack Command Integration: /get_feedback_insights last_month returns live metrics from Tableau.

How We Built It

  • Defined custom Salesforce objects: Feedback, Cluster, ClusterMember, Score, AuditLog.
  • Built ingestion pipelines (Slack, forums, CRM or CSV mockups).
  • Applied sentence embedding + agglomerative clustering for NLP grouping.
  • Designed an explainable PriorityScore formula blending frequency, recency, and VIP weight.
  • Streamed everything to Data Cloud → Tableau Next dashboards.
  • Ensured privacy and governance with pseudonymized data and audit trails.

Impact

  • PMs cut hours of manual triage.
  • Leadership gets defensible, transparent prioritization.
  • Support teams see feedback visibly influencing roadmaps.
  • Decisions shift from “what’s loudest” to “what matters most.”

Why It Wins

  • End-to-end automation: Slack → Salesforce → Data Cloud → Tableau.
  • Enterprise-aligned stack: Uses governed Salesforce data, no exports or silos.
  • Explainable AI: Transparent formulas, not black boxes.
  • Instant value: AI insights directly in Slack and Tableau Next.

Next Steps

  • Live Slack and Community ingestion.
  • Sentiment analysis and persona-based feedback.
  • Automated Jira integration for roadmap updates.

North Star

Make it obvious what to build next — and why. Turn noise into narrative, opinions into evidence, and backlog chaos into confident, customer-anchored plans.

Share this project:

Updates