Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Most AI deployment policies stop at approved chat interfaces. Meanwhile, employees install browser copilots, AI extensions, and third-party plugins that never touch Microsoft's management stack. IT can't configure what it can't see, and Group Policy and Intune only govern Microsoft's world. This post covers what actually happens once AI tools show up outside policy, five things most teams miss, and how PolicyPak enforces controls directly on the apps and browser extensions themselves.

Peak season brings challenges to the hospitality industry every year. Thousands of guests, temporary staff, vendors, and business partners interact daily with reservation systems, management platforms, mobile apps, and loyalty programs. That operational complexity makes hotels a particularly attractive target for cybercriminals. Artificial intelligence hasn't created a new problem for hotels, it is simply accelerating an existing one: identity-based attacks.

Cursor and Claude Code now read source code, install packages, and push commits with much of the access a senior engineer has, and often with less oversight. Give an agent a prompt to fix a bug, and it may pull a private API key from a config file, pass a customer record into its context window, or send a snippet of proprietary logic to a third-party model provider to reason about the fix. Security teams built policy for developers typing code by hand.

Every SaaS tool your organization onboards now carries a hidden layer of AI risk. The chatbot on your CRM, the transcription service your sales team runs, the code assistant embedded in your IDE. Each one processes company data through models you did not build, in ways your vendor questionnaire was not written to catch. Traditional third-party risk management was designed to evaluate infrastructure, access controls, and data handling.

Author: Alexander Ivanyuk, Senior Director, Technology For many MSPs, AI still looks like an extra tool outside the real workflow: open a chatbot, paste in a ticket, ask for help, copy the answer back and continue working. It may save a few minutes, but it also creates more steps and more places where context can be lost.

Monitoring AI agents in production is a fundamentally different problem from monitoring traditional software or even generative AI models. Because agents run autonomously, chain multi-step reasoning across tools and systems, and change behavior as their underlying models evolve, standard software metrics like uptime and CPU utilization miss almost everything that matters. ‍