AI-rewritten: This is a summary of an article from IEEE Spectrum, rewritten by AI (Qwen, running locally) to make it easier to read. The facts come from the original article – read it for the full story.

---

The author, who previously worked in product management and data analytics at companies like Best Buy and Target, now leads enterprise AI transformation at Lowe’s. He argues that the goal is not to sell artificial intelligence but to use it for useful expertise when customers need it. This shift requires moving from building systems to governing them, defining how they operate, which decisions they can make autonomously, and which actions remain off-limits.

He coauthored a book called The Enterprise Brain, where he describes this change as the "governor shift." A 2025 report from MIT Media Lab’s Project NANDA found that despite an estimated US $30 billion to $40 billion in enterprise generative-AI investment, only about 5 percent of integrated pilots are generating substantial value. The author calls this pattern the "GenAI Divide," noting that companies often lack people who can direct these systems and stand behind the results.

To address these challenges, he offers six guidelines. First, recognize when you have become "human middleware" by pulling data between tools, a role AI agents now handle better but without judgment. Second, trade rules for principles to guide systems through unexpected situations. Third, write culture into code as machine-readable instructions organized into a constitution, doctrine, and playbook. Fourth, install a trust thermostat rather than a switch, using confidence scores to decide when humans must intervene. Fifth, fix context before governing by ensuring the system sees the whole picture through a process called CCRAG. Finally, learn to lead by exception, focusing only on ambiguous or high-stakes cases while routine workflows run automatically.