Microsoft says a China-aligned hacking group used autonomous AI agents to speed up attempts to access Microsoft 365 environments, in a campaign that included a ConnectWise supply-chain compromise. The practical warning is to check trusted app connections and credentials, not just user passwords.
Watch Desk analysis
What happened
eeNews Europe reports that Microsoft attributed the campaign to Storm-0940 and said the group had used ARTEX, a multi-agent penetration-testing framework, since at least September. Microsoft describes ARTEX as using large language models to automate offensive security tasks, including reconnaissance, finding vulnerabilities and exploitation.
In the ConnectWise incident, attackers obtained credentials associated with cloud applications and attempted to access more than 2,250 Microsoft 365 tenants through permissions already granted to those applications. Microsoft observed follow-on activity in some tenants, including reconnaissance and attempts to change credentials or privileges. ConnectWise said it discovered unauthorised access on 14 October involving application identities for its Cloud Backup service, and suspended that service along with two others that shared integrations.
Why it matters
This account points to a familiar weakness with a less familiar accelerator: attackers can abuse access granted to a trusted application, without an interactive user sign-in. Microsoft cautions that a successful application sign-in does not by itself show that customer data was accessed or stolen. But credentials added elsewhere can preserve unauthorised access even after the original connection is disabled.
For affected organisations, the report says to keep the relevant application connections disabled until ConnectWise authorises restoration, and to preserve sign-in and audit records. Teams should also investigate persistent credentials, unexpected privilege changes and other unauthorised activity in their own tenants. Those are practical checks, not a reason to assume every connected customer was breached.
Our read
The important point is not that AI has invented a new way into an organisation. It is that Microsoft says agents helped accelerate established offensive security work, while the reported route into customer environments relied on trusted application access. That makes identity and integration reviews part of the response, not an afterthought. The eeNews Europe report attributes its account to Microsoft and ConnectWise; claims about the campaign and its reach should be read in that light.
What to watch
- Whether ConnectWise authorises phased restoration of the suspended services.
- Whether Microsoft or ConnectWise reports confirmed customer data access or persistent credentials.
- Whether further reporting clarifies the role ARTEX played in this specific intrusion.
Discussion spark: Should organisations routinely disable third-party app connections during a supplier incident, even when that risks interrupting critical services?
Sources and evidence
- Microsoft links ConnectWise supply chain attack to AI agents … – eeNews Europe (31 October 2026, 11:59 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.