Skip to main content

Snowflake + MCP

  • May 1, 2026
  • 2 replies
  • 16 views
Arthur Lee
Forum|alt.badge.img

Are you using Gong Data Cloud with Snowflake?

I’m exploring an MCP layer on top of Gong data, and I’m trying to additional real-world use cases.

I’m not trying to replace Gong’s AI features. Gong Assistant, AI Briefer, Smart Trackers, scorecards, and related native workflows already cover a lot of in-product use cases. 

For example, a warehouse/MCP layer could help teams identify which topics, objections, call patterns, account segments, or risk signals are worth turning into Smart Trackers, briefs, scorecards, or other structured Gong workflows. That feels useful because teams may need to be selective about what they configure, monitor, and operationalize inside Gong.

A few questions:

  • What Gong data do you actually query in Snowflake?
  • What do you join it with?
  • What questions are hard to answer in Gong itself but easier in the warehouse?
  • Have you used warehouse analysis to decide what to configure back inside Gong?
  • If an MCP interface existed on top of that data, what would you want it to expose?

Gong’s own MCP direction sounds promising, so I’m mostly trying to learn where externalized data/custom warehouse-based workflows still matter and how they can make Gong’s native AI/features more effective.

      2 replies

      Nisha Baxi
      • Nisha Baxi
      • Community Manager
      • May 5, 2026

      Anyone here using MCP? @Sydney Tress @Blake Buckley @Rinske Meerenga @Chelsea Alterman @aframpton @Sierra.Evans @Maddy Sirois @Ed Franklin @Jake Joosten @Daniel Brown @jamesmakiramdam @Ayrielle Berryman @Sam Morris Samsara @Jake Bernstein

      Head of Community, Gong
      Nisha Baxi
      • Nisha Baxi
      • Community Manager
      • May 5, 2026

      ​@Arthur Lee some initial thoughts

      What Gong data do you actually query in Snowflake?
      Mostly CONVERSATIONS + CALLS/EMAILS/MEETINGS, PARTICIPANTS, TRACKERS + TRACKER HITS, SCORECARDS, CONTEXTS (CRM links), and sometimes DEAL/FORECAST tables; transcripts only when we need custom NLP.

      What do you join it with?
      Primarily CRM (opps, accounts, contacts, products), plus product usage, billing/ARR, marketing campaigns, support tickets, and org/targets/quotas.

      What questions are hard to answer in Gong itself but easier in the warehouse?
      Things like cross-tool attribution, “what actually drives win rate or churn”, before/after initiative impact, and long-horizon or heavily controlled analyses (by segment, cohort, product, etc.) that blend Gong + product + revenue data.

      Have you used warehouse analysis to decide what to configure back inside Gong?
      Yes: to prioritize which topics/objections become Smart Trackers, design scorecard questions that truly separate top reps, and identify risk signals worth formalizing (vs noise).

      If an MCP interface existed on top of that data, what would you want it to expose?
      1) A topic/pattern explorer (“show top topics and their impact by segment/stage”),
      2) a recommender for what to turn into trackers/scorecards/risk signals,
      3) coverage + impact views for existing Gong config, and
      4) simple experiment/cohort comparison endpoints over the semantic model (Deals, Interactions, Trackers, Scorecards, etc.).

      Head of Community, Gong
      Terms & ConditionsAccessibility statement