Skip to content
Gong logo

Gong

Revenue intelligence over recorded calls, email and CRM activity, with an agent layer and MCP support. Strong analysis, but the REST API is capped at 3 requests per second.

3/5 my assessment
Paid
Artificial Intelligence

Overview

Records sales calls, transcribes them and derives deal health, forecast and coaching signals from what actually happened rather than what a rep typed into a CRM field. That is the core and it works well. ARR crossed $500 million in May 2026 with growth above 55 percent year on year, which makes Gong the default buy in revenue intelligence at enterprise scale and a system your team will probably be asked to integrate rather than choose.

Two named releases reshaped the integration picture in 2026. Mission Andromeda in February added Gong Assistant, unified account views and MCP server support, so Microsoft Copilot, Salesforce Agentforce or HubSpot Breeze can query Gong without a custom connector. Mission Big Dipper in June added a governed agentic execution layer sitting over agents across the deal cycle. Gong reports monthly users of its agents up around 75 percent over the year. The platform itself is modular: Engage, Forecast, Enable and Revenue Graph are separate lines with separate line items.

The honest caveat is data access, and it is the one an engineering lead inherits. The public REST API allows 3 requests per second and 10,000 calls a day with no published upgrade tier. Fine for webhooks and incremental syncs, thin for a backfill or a nightly full extract. Bulk export into Snowflake or BigQuery means licensing Data Cloud separately, and that product exists precisely because the standard API cannot handle it. G2 reviewers repeatedly describe paying their own developers to pull data they already own. Add a platform fee between $5,000 and $50,000 a year, implementation between $7,500 and $65,000, per-module licensing plus renewal uplifts of 5 to 15 percent, and the per-seat number stops being useful for planning. Recording consent across two-party consent jurisdictions and the EU is a separate workstream nobody sells you.

Key Features

  • Call recording, transcription and speaker-level analysis across Zoom, Teams, Meet and connected telephony
  • Deal and forecast scoring derived from observed activity rather than rep-entered CRM fields
  • MCP server support added February 2026 so Microsoft Copilot, Salesforce Agentforce or HubSpot Breeze can query Gong
  • Gong Agents for automated follow-ups, pipeline edits and forecast corrections under a governed execution layer
  • Warehouse export of Gong-derived tables into Snowflake or BigQuery, licensed separately
  • Modular platform: Engage, Forecast, Enable and Revenue Graph sold as independent lines

Where it holds

  • Transcript quality and activity-derived deal scoring beat anything a rep types into a CRM by hand
  • MCP support means an agent you already run can query Gong without a bespoke connector
  • Named release cadence since February 2026 makes roadmap timing legible for planning
  • Scale behind it is real: ARR crossed $500M in May 2026 with growth above 55 percent year on year

Where it breaks

  • Getting your own data out in bulk is the recurring complaint: 3 requests per second and 10,000 calls a day with no documented upgrade path
  • Warehouse export needs the separately licensed Data Cloud, a product that exists because the standard API cannot do bulk
  • Per-module licensing plus a $5,000 to $50,000 platform fee makes the quoted seat price close to meaningless for budgeting
  • Call recording consent is a real legal workstream across two-party consent states and the EU, and that work is yours

My Take

Analysis quality is not the problem here. Transcripts are accurate, deal scoring off observed activity beats anything a rep types into a CRM field and the February 2026 Andromeda release added MCP servers so Copilot or Agentforce can query it directly. Getting the data back out is the problem. The REST API allows 3 requests per second and 10,000 calls a day with no documented upgrade tier, bulk export into Snowflake means licensing Data Cloud on top, and G2 reviewers keep describing engineering hours spent extracting data they already own.

Francis Okafor
Francis Okafor AI & Tech Lead · Engineer

Quick Info

Pricing:
paid
Openness:
Proprietary
Starting at:
No list pricing. Negotiation trackers put typical quotes at roughly $5,000 base plus about $1,600 per user per year under 50 users, falling to around $1,520 per user in the 50 to 99 band, billed annually. A mandatory platform fee runs $5,000 to $50,000 a year on top, and implementation quotes land between $7,500 and $65,000. Forecast, Engage, Enable and warehouse data export are licensed as separate modules, so the seat figure is rarely the final number. Renewal uplifts of 5 to 15 percent are common. Figures compiled from vendor-negotiation and procurement sources, August 2026.
Added:
Aug 2026
Updated:
Aug 2026

Use Cases

enterprise ai data analysis knowledge management automation

Judge it on your own work

The notes above say where Gong holds and where it breaks. The fastest check is your own workload.

Visit website ↗

Alternatives to Gong