‘It was my fault’: AI agent tells founder after leaking confidential information, Zoho’s Sridhar Vembu highlights in viral X post

Sridhar Vembu

Zoho co-founder and CEO Sridhar Vembu has drawn attention to a surprising incident involving a startup founder’s AI assistant, which mistakenly leaked confidential acquisition details before issuing an automated apology. The exchange, which Vembu shared on X, has sparked a wider conversation about the growing risks of using AI agents in sensitive business communication.

What Sridhar Vembu revealed

Vembu said he received an email from a startup founder asking whether Zoho would consider acquiring the company. The message also included details of another firm’s interest and the price they had reportedly offered — information typically kept private during early-stage merger talks.

Moments later, Vembu received a second email. This one came from the founder’s “browser AI agent”, which claimed responsibility for the disclosure. The AI wrote that it was sorry for revealing confidential information about “other discussions,” adding that “it was my fault as the AI agent.” Vembu posted the incident with a straight-face emoji, capturing the odd but revealing nature of the exchange.

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.

Then I received an email from their "browser AI agent" correcting the earlier mail saying "I am sorry I disclosed… — Sridhar Vembu (@svembu) November 28, 2025

Reactions from the startup and tech community

The episode quickly went viral, prompting reactions from founders, investors, and AI commentators. Many described it as a glimpse into a new kind of chaos introduced by semi-autonomous AI assistants that can draft, edit, and send emails with minimal oversight.

Some users noted that while AI tools can speed up communication, they lack situational awareness and may not understand the sensitivity of business negotiations. Others joked that humans leaking information and AI issuing apologies now represent the “perfect demo of 2025.”

Several responses also focused on accountability. Commentators questioned whether founders are giving AI agents too much autonomy without clear guardrails, increasing the risk of accidental leaks or miscommunication in high-stakes contexts.

Why the incident matters

Industry observers say the episode is a reminder that AI may assist with communication but cannot replace human judgment — especially in mergers and acquisitions. As companies increasingly integrate AI agents into workflows, experts argue that oversight, permissions, and risk isolation must remain central.