Salesforce Integration for VoC

How Salesforce fits into a Voice of the Customer agent system — bidirectional sync, field mapping, and the tools Ramp uses to bridge CRM and customer intelligence.

Role of Salesforce in VoC

• Primary system of record for deals, accounts, contacts, and pipeline
• Enriches customer feedback with deal context — stage, ARR, segment, owner
• Enables routing insights back to account teams via CRM fields and alerts

Integration Approaches

1. Direct Agent Connection (LangChain)

• Ramp's LangChain agents connect directly to Salesforce alongside Gong, BigQuery, Apollo, Gmail, LinkedIn, and Exa
• Agents can query opportunities, contacts, account history in real-time
• Best for: dynamic queries during research tasks

2. Actively.ai (Intelligence Layer)

• Reads both standard and custom Salesforce fields
• Combines CRM data with Gong transcripts + third-party signals
• Outputs: account scoring, prioritization, territory optimization
• Provides "why you, why you now" hypotheses grounded in CRM + call data
• Best for: sales-led VoC where account context matters

3. Momentum.io (Revenue Ops Automation)

• AI extracts details from call transcripts and auto-populates Salesforce fields
• Generates AI summaries for sales-to-CS handoffs: deal history, champions, objections, next steps
• Creates automated Slack deal rooms with cross-functional visibility
• Best for: keeping CRM data fresh without manual entry

4. Gong Bidirectional CRM Sync

• /crm/object endpoints enable pushing call outcomes back to Salesforce
• /crm/map-fields configures which Gong fields map to which Salesforce fields
• Push: call summaries, next steps, competitor mentions -> Salesforce
• Pull: opportunity data, deal stage -> Gong for enriched call context

Data Available from Salesforce for VoC

Object	VoC Signal
Opportunity	Deal stage, close date, amount, win/loss reason
Case	Support issues, resolution time, satisfaction
Account	Segment, industry, ARR, health score
Contact	Champion vs. detractor, role, engagement
Task/Activity	Meeting frequency, email cadence, responsiveness
Custom fields	Product usage tier, feature requests, NPS score

Architecture Pattern

[Salesforce] <--bidirectional--> [Gong]
      |                              |
      v                              v
[Snowflake CDP] <--- dbt transforms
      |
      v
[VoC Agent] queries unified data
      |
      v
[Hightouch] pushes insights back to Salesforce

The key insight: Salesforce is both a source (deal context) and a destination (insights pushed back to account teams).

Key Takeaways

• Salesforce provides the deal and account context that makes raw customer feedback actionable
• Use multiple integration patterns — direct agent queries for research, Actively.ai for scoring, Momentum.io for keeping CRM fresh
• Bidirectional sync is critical — insights must flow back to where sales teams actually work
• Hightouch (reverse ETL) closes the loop by pushing computed insights back into Salesforce fields
• Don't build custom CRM integrations from scratch — the middleware ecosystem (Actively, Momentum, Hightouch) is mature

Related

• Gong Integration
• Architecture Patterns
• GTM Agent Economics