AB-731: AI Transformation Leader

Generative AI creates new content and differs from traditional AI, which analyzes data for fraud detection, spam filtering, and image recognition, using prompts to generate code, text, and images.

Learn how large language models predict the next token from context and why they do not truly know content, and how neural networks and transformers process prompts into code.

Explore Gen AI use cases for business, including customer support and data analysis, and map needs to Microsoft AI services like CallPilot, CallPilot Studio, Foundry, Security CallPilot, and GitHub CallPilot.

Describe the core differences among five AI model types: large language models, code models, diffusion models, multimodal models, and domain-specific models, and their use cases and examples.

Explore the main cost drivers of generative AI, including people, licensing for M365 Copilot, Copilot Studio, and Foundry, plus pay-as-you-go, tokens, and related Azure infrastructure.

Identify challenges of generative AI, including hallucinations, reliability, and bias, and outline grounding with retrieval augmented generation and human review to ensure data quality and fair outcomes.

Describe how prompt engineering shapes interactions with generative AI by defining roles, tasks, context, and output formats. Experience reduced hallucinations and improved accuracy, relevance, consistency, and efficiency.

Explore retrieval-augmented generation for AI, which finds real-time information and augments knowledge. Leverage live sources such as Python docs, GitHub, and Reddit to generate responses and reduce outdated data.

Identify core AI risks such as over-reliance, hallucinations, bias, misinformation, and disinformation, and examine cybersecurity threats like prompt injection and data disclosure in enterprise contexts.

Explore how machine learning merges data science and software engineering to build predictive models, with uses like forecasting and anomaly detection, and Gen AI as a subset of ML.

Define tasks and success metrics for a business problem, collect and prepare data, train and validate models, deploy, monitor with MLOps, and continuously improve the machine learning lifecycle.

Explore Azure Machine Learning, an end-to-end lifecycle platform that accelerates training, deployment, and model management with built-in MLOps, code-first and no-code options.

Map business processes to Copilot by exploring how M365 Copilot, integrated with Microsoft Graph, grounds prompts, ensures compliance, and enhances productivity across Outlook, Word, Excel, Teams, and more.

Explore how to access and use the Microsoft 365 Copilot chat on copilot.microsoft.com, sign in, choose enterprise vs consumer, and prompt blockchain in simple terms.

Understand agentic AI by comparing it to GenAI, and explore autonomous and end-to-end tasks handled by coding agents, support agents, research agents, and workflow automation across systems.

Explain how Copilot is a user-facing, reactive assistant, while agents operate autonomously to execute end-to-end workflows with tools and APIs.

Create agents in M365 Copilot by naming, selecting a subscription, and setting instructions and knowledge sources. Use templates like career coach to prefill goals and prompts, then deploy and interact.

Compare free consumer Microsoft Co-Pilot, Microsoft 365 base, and the Microsoft 365 Co-Pilot add-on, detailing chat, web data, work data, agents, and admin controls.

Learn to assign the correct Microsoft 365 Copilot licenses in your environment, distinguish add-on versus basic licenses, and follow Microsoft licensing articles to configure them in the admin center.

Explore how Microsoft Copilot Studio enables non-coders to build AI agents and agent flows with a low-code interface, connecting APIs, webhooks, and end-to-end workflows.

Create a Copilot Studio agent by starting blank or defining via natural language, then add knowledge from a white paper and test before publishing with authentication controls.

Explore when to build agents in the M365 Copilot Agent Builder versus Copilot Studio, using a simple decision diagram that weighs external scope, cross-team needs, complex workflows, and granular controls.

Explore how the Microsoft Graph API grounds Microsoft 365 Copilot by using user identity, devices, installed apps, and organization data to tailor responses and ensure compliance.

Demonstrates how Word Copilot can draft a cybersecurity engineer job description—five years of experience, CISSP, Azure, Python, and PowerShell—via pencil prompts and chat.

Learn to draft emails with Copilot in Outlook, tailor tone, rewrite and coach content, then customize drafts and use Copilot to summarize or prioritize emails.

Map business processes to Microsoft AI services such as CoPilot, CoPilot Studio, and Foundry, including M365, Security, GitHub, and Azure integrations.

Explore researcher and analyst agents in M365 Copilot for document summarization, insights, trend detection, and KPI forecasting. Build custom agents without code to automate processes, connect APIs, and end-to-end workflows.

Discover Azure's global backbone, regions, availability zones, and region pairs to reduce latency, ensure data residency, and enable resilient multi-data-center deployment.

Explore Azure regions, data centers, and the Azure backbone, highlighting availability zones, pad regions, geography, and key compliance like GDPR, HIPAA, and PCI DSS.

Explore the Azure resource hierarchy, from root management group to subscriptions, resource groups, and resources, and apply governance with policy and RBAC for secure, cost effective workloads.

Show how Entra ID tenants manage identities (human, managed, and application registrations) separate from Azure resources like subscriptions, management groups, resource groups, virtual machines, virtual networks, and app services.

Learn how to create an Azure subscription by starting free or pay-as-you-go, with 12 months of popular services, always free options, and a $200 credit for 30 days.

Explore the Azure resource hierarchy, including management groups and subscriptions, rename and move subscriptions, and create a demo group and resource group in North Europe.

Create and scope an Azure cost management budget for a subscription, set monthly (or quarterly) periods, thresholds, and alerts for actual or forecasted spend—without stopping resources.

Discover Microsoft Foundry, a unified enterprise AI platform for building and orchestrating multi-agent workflows with Azure integration, model catalogs, and security.

Explore foundry models, from fast, pre-built tools to foundation models, RAC grounding, and fine-tuned custom models, and learn when to use each for accuracy, creativity, and specialized workflows.

Explore how the Foundry agent service functions as an assembly line for building agents, converting inputs into LLM-driven outputs using tools, memory, multi-agent orchestration, and RBAC-enabled security.

Learn how Foundry IQ adds a unified knowledge layer for agents with multi-hop retrieval, source routing, and permission-aware access across data sources like One Lake and SharePoint.

Explore Microsoft Foundry and its tools, such as translator, vision, search, speech, language, and documents intelligence, and how these tools function as specialized APIs alongside the agent service and models.

Run Microsoft foundry models locally with foundry local, including service, runtime, cache, and model management. Interact via http or named pipes with your sdk for on-device npu gpu cpu inference.

Explore Microsoft responsible AI and its principles: fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability. Apply prompt safeguards and identify-measure-mitigate-operate loop across Copilot, Copilot Studio, and Boundary.

Explain the importance of responsible AI and leadership accountability, addressing design, ethics, privacy, and the evolving defenses needed to mitigate bias, risks, and unintended consequences.

Establish governance for AI use in your enterprise by applying the Microsoft responsible AI principles—fairness, reliability and safety, privacy and security, inclusiveness, transparency, accountability, and apply them in daily work.

Implement responsible AI principles across the lifecycle with a hybrid governance model led by a chief AI ethics officer, an AI ethics office, and an AI ethics committee.

Identify sensitive ai workloads with a sensitivity flag, assess impact and risk, and apply mitigation under the ai ethics office to govern ai responsibly at scale.

Plan for AI adoption guides organizations to achieve business impact by embracing adoption mindset and following a four-step framework: educate, readiness, map the journey, and start building the agentic future.

Align leaders around the company's top priorities to solve core business problems with AI, through executive envisioning, focused sessions, industry scenarios, and two to three strategic bets.

Assess your AI readiness with a secure evaluation across five dimensions—business strategy, technology strategy, AI experience, organization and culture, and AI governance—mapping maturity to identify early risks and guardrails.

Establish a center of excellence to map AI journey with intake, prioritization, and skills development, and align governance with responsible AI principles through data and platform landing zones and MLOps.

Identify and prioritize high-value AI opportunities through discovery workshops with business owners, evaluate use cases for impact and feasibility, and decide to buy, extend, or build to avoid pilot sprawl.