Tool LaunchGongMission Big DipperGong Revenue HarnessCustom AgentsGong Assistant

Gong Launches Mission Big Dipper and the Revenue Harness Agentic Layer

Gong's Mission Big Dipper debuts the Revenue Harness, a governed agentic execution layer, plus generally available Custom Agents for RevOps and sales.

5 min readPublished 2026-06-27

What happened

On June 24, 2026, Gong announced Mission Big Dipper, the latest product mission for its revenue intelligence platform, introducing a new agentic execution layer and expanded capabilities across the Gong Revenue AI Operating System. The centerpiece is the Gong Revenue Harness — described as the natural evolution of Gong's prior investments in Agent Studio and Model Context Protocol (MCP) support — a layer that governs, orchestrates, and connects AI agents across the full revenue cycle.

Gong positions the launch as a response to a specific failure mode in enterprise AI: generic agents that break in production because they lack revenue context, governance, and human oversight. "Almost none of them are moving the number with it," said Eilon Reshef, Gong's Chief Product Officer and co-founder, referring to revenue leaders who have adopted AI but struggle to show measurable results. "The Gong Revenue Harness gives agents the context to reason over real customer conversations, the permission models revenue teams require, and blueprints reverse-engineered from your actual wins."

The most prominent new capability is Custom Agents, now generally available. Custom Agents let any RevOps leader or sales manager describe a workflow in natural language inside Gong — for example, monitoring accounts over $100K for specific risk signals and alerting the account executive before Monday morning — without engineering support. Because these agents run on the Revenue Harness, Gong says they operate with strict enterprise control, scoped data access, a full audit trail, and configurable human oversight.

Why it matters for practitioners

Mission Big Dipper reframes the agentic AI conversation around governance and outcomes rather than raw capability. For revenue and competitive teams, that shift carries several practical implications.

1. Governance becomes the differentiator, not the model. Most vendors can now wire a large language model to CRM data. What stalls enterprise deployment is trust: scoped permissions, auditability, and a human in the loop. By making the Revenue Harness the substrate that every agent runs on, Gong is betting that governed execution — not model novelty — is what converts AI pilots into production. Teams tracking Gong's competitive positioning should note that the company is competing on control-plane maturity, an axis where incumbents with deep data and security investments have an advantage.

2. No-code agent building moves authority to the front line. Custom Agents put agent creation in the hands of RevOps leaders and sales managers rather than engineers. This compresses the build-to-deployment cycle and lets revenue teams encode their own deal intelligence logic — risk signals, account thresholds, escalation rules — without a development queue. The operational consequence is that competitive and revenue workflows can be iterated weekly instead of quarterly.

3. Enablement folds into the agentic layer. Gong is extending sales enablement capabilities into the same agentic framework through tools like AI Coach and AI Builder for Scorecards, signaling that rep readiness, coaching, and execution are converging onto a single operating system rather than living in separate point products.

Key details

  • Announcement date: June 24, 2026
  • Product mission: Mission Big Dipper, the latest mission for the Gong Revenue AI OS
  • Core launch: Gong Revenue Harness — a governed agentic execution layer that orchestrates and connects AI agents across the revenue cycle
  • Lineage: Built on Gong's prior Agent Studio and Model Context Protocol (MCP) investments
  • Custom Agents: Generally available this month; natural-language agent creation for RevOps and sales managers with no engineering required
  • Governance features: Strict enterprise control, scoped data access, full audit trail, configurable human oversight
  • Also GA this month: AI Builder in Gong Assistant, Gong Assistant as a standalone workspace, AI Coach, and AI Builder for Scorecards
  • Coming July 2026: Gong Assistant in Account Console and Dry Run
  • Executive: Eilon Reshef, Chief Product Officer and co-founder

Market implications

Mission Big Dipper intensifies the agentic-AI race among revenue platforms. Gong is arriving at the governed-agent conversation from a position of strength — it disclosed surpassing $500M ARR with growth past 55% year-over-year earlier in 2026 — and is now using that data and distribution advantage to define what "production-ready" agentic revenue AI looks like. The framing is deliberate: by emphasizing audit trails, scoped access, and human oversight, Gong is targeting the trust barrier that the company's own research has identified as the primary reason enterprise AI projects stall.

The competitive context is crowded. Salesforce has pushed Agentforce multi-agent orchestration to general availability and acquired Momentum to bring native conversation intelligence into its CRM, while Clari and Salesloft continue building out their own agentic revenue systems. Against that backdrop, Gong's Revenue Harness is a claim to own the orchestration layer — the part of the stack that decides what agents can do, with which data, under whose supervision. Whether that claim holds depends on how cleanly Custom Agents scale across large customer bases and whether the governance model genuinely reduces deployment friction relative to rivals.

For practitioners in the revenue intelligence category, the strategic takeaway is that the evaluation criteria are shifting. The question is moving from "can this platform generate insights?" to "can it safely act on them at scale, with auditability my security team will accept?" Teams comparing vendors should pressure-test the governance layer directly — permission granularity, audit completeness, and human-in-the-loop controls — rather than evaluating agent capabilities in isolation.

Related resources