Your team is using AI tools you don’t know about.
Not because they’re hiding anything — they just didn’t think it was a big deal. A developer added Context.ai to their Vercel workflow. A marketer signed up for an AI writing tool with their work email. Someone’s running a chatbot prototype on a free tier account.
Then Vercel got breached. Through Context.ai.
What Actually Happened
Vercel — a major web infrastructure company trusted by thousands of businesses — disclosed a security breach in April 2026. The attack path: an employee’s account was compromised via a third-party AI tool they’d been using. The attackers accessed internal systems and grabbed customer credentials.
No zero-days. No super-sophisticated hack. Just a SaaS AI tool nobody in security reviewed getting compromised, and boom — access to a major tech company’s internal environment.
ShinyHunters, the same cybercriminal group behind the ADT breach exposing 10 million records, claimed responsibility.
This Is the Attack Pattern Nobody Is Talking About
Here’s what’s keeping cybersecurity teams up at night right now: your attack surface isn’t just your firewalls and servers. It’s every AI tool your employees have adopted — often without procurement, IT approval, or security review.
You’re managing M365. You’re managing your network. But are you managing the AI tools your team connects to those systems?
The answer for most small and mid-sized businesses is: no.
And attackers know it. China’s-nexus cyber actors have already shifted tactics to hiding inside “covert networks” of compromised tools. Hospitals and critical infrastructure are targets. But the same logic applies to any business: if your employees are connecting unvetted AI tools to your data, that’s an open door.
What You Should Do Now
-
Audit your AI tool sprawl. What has your team connected to your M365, Salesforce, cloud environments? If you don’t know, that’s the problem.
-
Treat AI tools like any other vendor. They need security review, not just adoption.
-
Get managed IT that thinks about this. Break-fix support doesn’t catch this. You need someone monitoring your entire stack — including the AI layer.
The businesses that’ll get burned this year aren’t the ones ignoring AI. They’re the ones adopting it faster than they’re securing it.
Want a security review that covers your AI exposure? Talk to NSI Tech. We’ll map your actual risk.