ZoomInfo launches Agent Teams: AI agents that run the whole sales play — built on yesterday’s DoubleO.ai buy

ZoomInfo announced today Agent Teams, a new product that hands revenue teams AI agents capable of running complete go-to-market plays end to end — from buying signal to outreach action — without a human babysitting every step. It is available now inside GTM Studio, and ZoomInfo says existing customers get it with no new SKU and no new contract.

The launch lands one day after ZoomInfo announced it had acquired DoubleO.ai, the orchestration startup founded by Derek Osgood and Karthik Suresh. Agent Teams is built on DoubleO's orchestration architecture — an acquire-and-ship cycle of roughly 24 hours that says plenty about the pace of the agentic GTM land grab.

From signal to action, unsupervised#

The pitch is aimed at a specific frustration: most AI in sales tooling today is fragmented — an assistant in the CRM, enrichment in another tool, sequencing somewhere else — and the research, follow-ups, and handoffs between judgment calls still land on people. RevOps teams stitch together close to a dozen systems to understand an account before they can act on it, while high-intent signals sit waiting.

Agent Teams coordinates teams of specialized agents across a company's systems, its buying signals, and ZoomInfo's own intelligence to execute full GTM motions. The company's go-to example: a former champion becomes CRO at a target account. A traditional workflow creates a CRM task and stops. Agent Teams checks whether the new company is already a customer and fits the ideal customer profile, reconstructs the champion's role in the original deal, maps the new buying committee, and drafts outreach grounded in that relationship. When the champion visits the pricing page two weeks later, a website-visit trigger picks the play back up with full context intact.

The architecture bet: shared context#

Under the hood, Agent Teams runs on what ZoomInfo calls a real-time evaluation architecture. Each specialized agent has its own role, instructions, and tools, but they share context from every prior step and decide the next best action from the current state of the account record — rather than following a rigid flowchart.

“Revenue doesn’t follow a flowchart. Champions change jobs, deals stall, and buyers come back months later,” said Henry Schuck, founder and CEO of ZoomInfo. “Rigid workflow automation breaks the moment that happens. Agent Teams works from what’s actually true about an account right now, and keeps the GTM play moving when things change.”

“Most AI agents exchange outputs, not understanding, so the reasoning behind each step gets lost at every handoff,” added Dominik Facher, chief product officer at ZoomInfo. “Agent Teams runs on a real-time evaluation architecture where agents share context and build on each other’s work, grounded in the richest GTM intelligence available.”

What teams actually do with it#

Beyond champion tracking, ZoomInfo describes plays like cloning closed-won deals: Agent Teams reviews the full sales cycle — call transcripts, emails, CRM records — to understand why the customer bought, drafts a case study, finds ten competitors of the new customer that aren’t customers yet, and reaches out to each with tailored messaging. The company cites Silverfort’s sales operations leader Caroline Burnett, who said the product let her “scale RevOps and accelerate our GTM both in quality and outcomes.”

Governance gets a paragraph of its own in the announcement: the product runs on ZoomInfo's data foundation with privacy-rights management, SOC 2 controls, and ISO 27001 and ISO 27701 certifications — the checkboxes enterprise buyers now demand before letting agents touch customer data and outbound messaging.

What it changes#

ZoomInfo started as a B2B data vendor. Agent Teams is the company's clearest move yet from selling intelligence to selling outcomes — and from a product its customers' RevOps teams configure into a system that configures itself. The DoubleO.ai acquisition closing into a shipped product within a day also signals how the agentic stack is being assembled right now: buy the orchestration layer, plug it into the data moat, ship.

The open question is the one every agentic-GTM vendor faces: whether agents that draft outreach and trigger plays across the buying committee earn trust or trigger spam filters — human and algorithmic. ZoomInfo's answer is shared context and real-time evaluation instead of brittle workflows. Revenue teams get to find out starting today, at no extra line item.