Microsoft Learn — searchable YouTube transcripts & AI Q&A

Trust is the infrastructure for AI speed | The AI-Ready Executive

Nathan Cook argues that trust is the foundation for AI adoption, and governance is the mechanism to build it. He explains that a simple governance policy, which initially approved only Copilot, drove a 98% uptake in the first year by funneling staff into a safe tool and enabling measurement of value. He emphasizes that governance must be flexible and continuously updated to keep pace with AI changes, and that involving people in the design process and ensuring AI is unbiased are key to maintaining trust.

Who actually owns data governance? | Ask a Cybersecurity Pro

Data governance should be owned by the business, but it only works as a team sport involving legal, compliance, security, and technology, with named accountable data owners. Governance sets the rules for information, while security enforces them, and both are necessary to avoid protecting data that shouldn't exist or having policies that don't hold up. To adapt to changing regulations and AI, build governance around durable principles like knowing what data you have, why you keep it, and who can access it, rather than creating separate programs for each new rule.

Where do you start with data classification? | Ask a Cybersecurity Pro

Data classification should start by identifying the most sensitive information that would cause the greatest impact if exposed, rather than trying to classify everything at once. The speaker distinguishes classification (the decision about data sensitivity) from labeling (the signal that communicates that decision to people and systems), and recommends using automation for consistency and scale while relying on employees for context. In the age of AI, unclassified data can be exploited at machine speed, so organizations must be deliberate about what information is accessible and how it is identified.

Azure Demo: Estimating Costs with Tokens

Estimating AI costs requires understanding token usage, which varies by model and efficiency, and using Azure Cost Management to analyze spending by service and meter. The speakers demonstrate how to identify cost spikes and trace them to specific model deployments in Microsoft Foundry, emphasizing that end-user behavior and model efficiency can significantly impact consumption-based costs. They caution that not all AI services are token-based, so it's important to know how each service is priced.

Find the Signal in the AI Noise | The AI-Ready Executive

Tristan de Montebello argues that leaders should use AI as a sparring partner to pressure-test their thinking rather than letting it do the writing, preserving authenticity and trust. He recommends reassuring teams by making AI concrete and addressing individual fears through one-on-one communication, while also using a 'north star' technique to find clarity amid noise. He emphasizes that imperfections and genuine communication are more effective for building trust than over-polished AI-generated output.