WorkOS says hackers used AI tool ARTEX to reach its sales records

WorkOS, which provides sign-in and identity services to software companies, has disclosed that attackers reached an internal tool used by its sales and customer-facing teams. Chief executive Michael Grinich said on Monday that the attack "involved ARTEX, an autonomous AI agent framework used to carry out offensive security operations", and that WorkOS's production systems were secure and no customer platform data was accessed.

The company found the unauthorised access on October 30, according to its security bulletin. It has disabled external access to the tool, hired an outside forensics firm and started a wider review of access controls on its internal tools.

What the attackers could see

The bulletin lists the data in the affected tool: business contact details such as names, email addresses, phone numbers and job titles; customer account histories and sales information; summaries and excerpts from sales calls, support discussions and customer Slack conversations; and operational data including product usage and "connection or directory health".

WorkOS says it is reviewing the records to work out the impact on individual organisations and will contact affected customers directly. It has not said how many customers are affected.

The bulletin itself does not name ARTEX. It says the investigation "identified the use of an AI-assisted offensive security framework to automate reconnaissance and exploitation attempts", and links to reports on attacks that used ARTEX by Wiz, Google and CrowdStrike.

What is not known

WorkOS has not said how the attackers first got into the sales tool, whether through a stolen credential, a third-party service or a software flaw, or who was behind the attack. ARTEX is an open-source framework that runs a team of AI agents against a target; at least four threat actors used it in intrusions documented in October, two of them tied to China, but no one has linked any of them to the WorkOS incident.

What customers should do

The data in the tool is the kind used to make phishing convincing: who a customer's contacts are, what they discussed with WorkOS and the state of their integration. WorkOS asks customers to "verify unexpected requests before sharing sensitive information or taking action", even when a message refers to a real colleague or a real support conversation.

WorkOS says it will publish further updates as its investigation confirms more.