Leaders are now expected to have a view on AI: which pitches to back, which claims to believe, which risks to watch, and what to tell staff who are already pasting work into chatbots. This free briefing follows Larkfield Home Services, a composite Southeast home-services company with a winter deadline, a $250,000 budget and a board meeting in January, through three acts: the problem, AI applied well, and AI applied badly and the fix.
You will sort proposals into predictive, generative and agentic AI and know the first question each raises; read a test on your own work instead of a vendor's accuracy claim; choose a first use case and measure it net of the time people spend checking AI output; name the risks that stay with you whoever builds the system; screen a vendor contract with six questions; and correct a 90-day plan and a one-page board summary. Every lesson works through real-looking material (a vendor term sheet, a 300-call test, a pilot readout, a chatbot transcript) and shows a failure and its fix. A short Decision Lab puts you in the COO's seat for the winter's three hardest calls (a chatbot's promises, a vendor's 24-month offer, the board page) and lets you live with the consequences. Real-world examples come from court and tribunal decisions, regulators and published studies, cited at the source.
The core path (six lessons with their checks, the lab and a 10-question scenario exam) takes about an hour; the exercises and the capstone, a two-page plan and board page, proposal screen or board review for your own job, with a template, are optional and come with model answers. Regulatory references are current as of October 2026 and are educational, not legal advice.