Time for our Big Ideas.
After 34 years at the helm, the Linux kernel community has established a contingency plan to replace its creator, Linus Torvalds. This decision comes as the community acknowledges that its leadership is aging, with Torvalds himself noting that the community is “getting grey and old.”
Microsoft is facing significant challenges with Windows 11 as user trust continues to erode due to ongoing bugs and performance issues. The operating system has encountered persistent problems such as intrusive ads and system prompts, leading to frustration among users. According to Pavan Davuluri, president of Windows and devices, the company is shifting its focus to address these pain points throughout the year, emphasizing improvements in performance and reliability. Microsoft acknowledges that trust has been compromised, particularly after controversial system requirements excluded many devices.
AI agents are rapidly transforming workplace productivity by automating a variety of tasks, from scheduling meetings to executing workflows in real time. However, as they gain traction, security teams are increasingly questioning, “Who approved this agent?” Unlike traditional users or applications, AI agents can be deployed quickly and often have broad access permissions, complicating accountability and traceability. These agents operate autonomously, acting on behalf of multiple users without ongoing human oversight, which can lead to significant security risks. For example, organizational agents, which are shared across teams, often lack clear ownership and accountability, heightening the potential for misuse and expanding their risk profile.
A recent survey from Drexel University’s LeBow College of Business reveals a significant gap between the rapid adoption of agentic artificial intelligence systems and the establishment of governance frameworks to manage them. While 41% of organizations now use these AI systems in their daily operations, only 27% report having mature governance policies in place to oversee their functioning.
Why do we care?
So here are the questions IT providers should actually be asking themselves.
Have you thought about your own contingency plans?
If a founder, a technical lead, or a key operator steps away—intentionally or not—does the business keep functioning, or does knowledge walk out the door with them? Open-source communities learned long ago that single points of authority don’t scale — many MSPs are still betting their businesses on exactly that.
When you recommend software, do you evaluate trust, or just capability?
Are you accounting for how vendor behavior—ads, forced upgrades, shifting requirements—erodes confidence for the people who rely on you to make good choices?
And on AI, let’s be blunt:
Do you know who approved the agents running in your clients’ environments?
Do you know who owns them?
Do you know who is accountable when they act?
Because governance keeps coming up in these stories for a reason. It’s not theoretical. It’s where authority, risk, and responsibility intersect.
Every one of those unanswered questions turns into non-billable labor when something goes wrong. And non-billable labor is how fixed-fee MSP models quietly collapse.
This is no longer just about managing systems. It’s about teaching clients how authority works in automated environments — because if you don’t define who’s allowed to act, you’ll be held responsible for what acts anyway.
And if you’re not asking the questions now, you’ll be answering them later—after something breaks.