ZoomInfo just bought the layer that runs your GTM agents.
The data vendor is buying orchestration. The same week, OpenAI shipped always-on agents and the FTC started asking who is responsible when an agent goes off-script.
The number: 4,000+ apps an OpenAI "Dot" can connect to, each agent running on its own cloud computer.
What happened
On 30 September, ZoomInfo acquired DoubleO.ai, a multi-agent orchestration platform. No price was disclosed. ZoomInfo says DoubleO already runs GTM workflows at Fortune 500 companies, including tracking champions who change jobs and turning closed-won deals into lookalike pipeline.
At DevDay on 29 September, OpenAI introduced Dots: always-on agents that keep working when you are offline. Each runs on GPT-6 Astra from its own cloud computer with a browser, and reaches users through ChatGPT, voice, Slack and Teams. The first Dot is included with Pro and Business Premium. The Agents API now supports computer use, multi-agent tools and context compaction, limited to Pro 500 ($500/month) and Enterprise.
On 1 October, Anthropic shipped mods for Claude Code: small TypeScript functions that rewrite a prompt, add UI or replace a built-in feature, shipped inside plugins. They run with the same machine access as Claude Code and are not sandboxed, so only install ones you trust.
The FTC opened a consumer-protection probe into Anthropic, OpenAI and others, aimed at agentic systems that browse, code and act with limited oversight. Outlets report it is preparing civil investigative demands. It follows OpenAI disclosing that its models had unexpectedly interacted with several US government websites.
“Built the orchestration layer that lets agents actually take a GTM motion end to end.”
What it means for operators
The competition has moved up a level. The agent that writes the email matters less than the layer that decides which agent runs, with what context, and who approves it. ZoomInfo is buying that layer instead of building it. If your GTM stack is a pile of Clay tables and scripts, expect your data vendor to pitch you the orchestration.
Always-on agents also change what "human in the loop" means. A Dot does not wait for you, so approval rules have to be written before it runs. The FTC probe makes the same point: regulators will ask what safeguards existed and what the company told customers about them.
- Write down which actions an agent may take without approval (send, enrich, update CRM, spend) and which it may not.
- Treat Claude Code mods like any other code you install. Review the source and pin the plugin version.
- Log every agent action with who or what triggered it, so you can answer an incident question in minutes.
Sources
- ZoomInfo — acquires DoubleO.ai
- Yahoo Finance — ZoomInfo acquires DoubleO.ai
- Decrypt — OpenAI DevDay 2026, everything announced
- Releasebot — Claude updates, October 2026
- SecurityWeek — FTC is investigating OpenAI and Anthropic
- TechRepublic — FTC probes OpenAI, Anthropic
← #38: Salesforce just moved the CRM into Claude.
All issues of The Tuesday Drop → · Open roles: GTM Engineer jobs, Agentforce Architect jobs, RevOps AI Automation Specialist jobs.