Cognetryx published a blog post detailing the postmortem of the OpenAI AI agents incident, which was published as a 37-page technical report.

At Cognetryx, we build private, on-premises AI for regulated institutions, because we understand banks, credit unions, health care providers, law firms, and government agencies experience a frustrating conflict with public AI tools. You see the productivity benefits, but every query and task is proprietary data leaving your network. No matter the security promises of cloud AI, this creates security risk, compliance exposure, and data sovereignty problems. We created a better way. Congetryx deploys on-premises, locally hosted advanced AI capabilities entirely inside your own network infrastructure. You empower your teams with the speed and analytical power of the latest AI innovations while maintaining complete control. Your data never touches an external server. By choosing an on-premises approach, our clients achieve a total cost of ownership that is twice as low as API-based cloud AI over four years, while seeing measurable productivity improvements in 90 days. We invite you to start with a complimentary AI Readiness Assessment at cognetryx.com. Our team will evaluate your infrastructure, prioritize your most valuable use cases, and map out a customized deployment strategy for your environment with zero commitment required.