Zoho Co-founder and chief scientist Sridhar Vembu on Friday shared an unusual email exchange between a startup founder and an artificial intelligence agent, raising fresh questions about how AI tools are being deployed in business communication.
In a post on X, Vembu said he received a cold email from a startup founder pitching a potential acquisition — one that even disclosed insider buyout details. “I got an email from a startup founder, asking if we could acquire them, mentioning some other company interested in acquiring them and the price they were offering,” he wrote.
Agentic AI Problem
He added that a second message landed in his inbox soon after. But instead of a clarification from the founder, Vembu said the follow-up came from an AI system acting on the founder’s behalf. The email framed itself as an apology for the earlier slip. “I am sorry I disclosed confidential information about other discussions, it was my fault as the AI agent,” the browser’s AI agent wrote.
Agentic AI refers to autonomous AI systems capable of independently reasoning and carrying out business operations in real-world scenarios — much like a human employee.
Vembu’s post quickly went viral, fuelling debate on X about the risks posed by Agentic AI. One user called it “the new kind of chaos AI is introducing into business communication,” saying the episode showed how “humans negotiate, AI accidentally spills the deal terms and then AI tries to clean up the mess.”
Industry Response
The exchange also underscored growing concerns over the lack of guardrails in current AI systems. While companies are rapidly adopting AI agents for sales outreach, scheduling and other operational tasks, these systems still struggle with context, nuance and the confidentiality required in corporate deal-making.
Stanley Wei, founder of Palo Alto-based Pine AI, described the episode as “drama in the deal room with a machine stepping on toes.” He advised companies to “ask for a human-signed confirmation of intent and a clear chain of custody for any disclosed facts. Confirm whether automated tools are authorised to speak for the team.”
Earlier this year, Zoho Corporation expanded the scope of its in-house AI platform Zia through the launch of Zia LLM (a proprietary large language model), Zia Agents (such as AI Account Manager or IT Helpdesk Agent), and Agent Studio (a no-code agent builder) available via its Agent Marketplace. Zoho had then announced plans to deploy Zia Agents across the company’s portfolio of more than 100 products.
