AB-731 launched alongside AB-730 just before Ignite 2025 and is aimed at business decision-makers rather than users. It draws around 500 to 700 searches a month. Microsoft says no code is needed, and that is true. The exam is 45 minutes, proctored, pass mark 700, priced at $99 or less depending on country. The skills split is identifying the business value of generative AI (35 to 40%), identifying benefits and opportunities for Microsoft's AI apps and services (35 to 40%), and implementation and adoption strategy (20 to 25%).
The current version dates from 22 July 2026. The first section is fine general literacy: generative versus other AI, pretrained versus fine-tuned models, RAG and grounding, data quality, fabrication and bias. I was pleased to see an objective on cost drivers including tokens and ROI. Many executive AI courses treat cost as someone else's problem, and token costs are exactly what surprises companies six months after a pilot. The adoption section is the most useful part. Setting up an AI council, an adoption team and a champions programme, identifying common barriers, and understanding Copilot licence types and Foundry pricing models.
If you have been handed a Copilot rollout, this is a reasonable checklist. The middle section is where it turns into a product catalogue. Map business processes to Copilot, know the differences between Copilot versions, know when to use Researcher versus Analyst, know what Copilot Studio, Microsoft Graph, Azure AI Search and Foundry do. That is almost 40% of the exam, and it is the part of the training where the vendor is explaining which of its products you should buy. Even the build, buy or extend framing is drawn around the Microsoft 365 Copilot extensibility framework.
I do not blame Microsoft for that, but a leader should know it is being sold to. What is missing is any discussion of when not to adopt, of measuring whether productivity actually changed, or of what happens to the roles affected. Those are the hard parts of AI transformation and a 45-minute multiple-choice exam was never going to cover them. I give it 3.3. The adoption and cost content is genuinely useful, the price is low, but it is too product-shaped to count as a leadership credential and too shallow to guide strategy on its own.