UNDERSTANDING AI By Christine Hopkins Protecting decision quality in the age of AI Most conversations around AI governance still feel strange to me because they tend to come from one or two extremes. Either the conversation sounds highly technical and compliance-driven, or it sounds overly optimistic in a way that does not match how most businesses actually operate day to day. For me, the experience has been much more practical than that. As the managing owner of both ASCI companies, my goal was never to turn us into an AI company. We were trying to figure out how to use these tools in a way that improved the business without creating new problems we did not fully understand yet. Ironically, governance itself was not the hardest part for us. The harder part was getting people comfortable enough with the tools to use them consistently and appropriately. I was using the tools heavily myself early on while also watching how people inside our company were experimenting with them. I was also talking to peers outside the company who were dealing with many of the same issues. Some people were hesitant to touch the tools at all, while others became overly dependent on them very quickly, especially for writing and communication. What became obvious was that adoption was happening faster than structure. I realized during that process that AI governance is not really about controlling the technology itself. It is more about protecting decision quality while people are moving quickly and while the tools are evolving around them. Early on, it is easy to think about AI as mostly a productivity conversation. At a certain point though, the bigger issue becomes operational consistency because these tools start influencing proposals, customer communication, policy writing, strategic discussions, pricing conversations, and internal decision-making. For me, one of the most important governance discussions became data protection. Very early on, we made the decision not to allow direct AI access into our Microsoft 365 environment. We intentionally chose not to connect SharePoint, Teams, or OneDrive directly into these systems while we were still learning how employees were using the tools and while we were still building standards internally. Instead, we created a separate controlled Google Workspace environment and moved only approved materials into that space for AI-accessible work. That approach was less convenient in some ways, but it gave us much more visibility and control over what information the systems could access and how people interact with it. At the same time, we also started realizing that familiarity with the tools does not automatically translate into good judgment when using them. I still regularly receive documents for approval where the writing sounds polished, but the reasoning has not been thought through carefully enough, or the document still carries that obvious AI cadence that people are starting to recognize immediately. Sometimes you can also see where content has been copied between systems, manually edited, reformatted, and pieced back together in ways that create inconsistencies that normally would not exist if someone had simply worked through the document carefully from the beginning. Early on, I would often just rewrite or clean those documents up myself because it was faster and easier. Eventually I realized that approach was actually preventing people from learning how to use the systems properly. Now I send many of those documents back with feedback asking employees to slow down, validate the reasoning more carefully, clean up formatting issues, and rewrite sections so they sound like a person instead of a generated summary. Otherwise, the organization slowly starts relying on polished language and speed instead of judgment and clarity. I also think women business owners sometimes underestimate how transferable our existing leadership instincts already are in this environment. Most of us are already balancing growth, operational realities, staffing challenges, customer relationships, financial pressure, and longterm reputation concerns at the same time. Those instincts are actually very useful when approaching AI governance because the conversation ultimately comes back to judgment, accountability, and risk management. AI can absolutely improve businesses and help people move faster, but leadership still has to own the judgment behind the work. CHRISTINE HOPKINS is President & CEO of Advanced Supply Chain International and ASCI Federal Services, Alaska-based companies providing commercial supply chain services and federal logistics support across the United States. Connect with Christine on LinkedIn at www.linkedin.com/in/ christineehopkins. Christine was a 2026 recipient of the Enterprising Women of the Year Award and is a member of the Enterprising Women Advisory Board. 50 enterprising Women witsarut sakorn / Shutterstock.com
RkJQdWJsaXNoZXIy MzA5OTEy