Lead the change
Choose a problem worth solving, assess readiness, compare investment options, set decision rights, and scale only when the evidence supports it.
Read the AB-731 strategy guideBusiness AI
From deciding where AI belongs to using it well at work. My business lens connects investment, people, and governance with practical, verifiable outcomes.
Each course has its own reading page: AB-731 for strategy and AB-730 for everyday use. Both contain original application notes. The practical track is no-code AI use, not developer training.
Start with the business decision
A useful answer is not yet a better process. Start with who benefits, what changes, and how the result will be checked.
Improve the work
Give AI a clear brief, use approved information, review documents and numbers, and keep control of sharing, scheduling, and delegated actions.
Read the AB-730 practical guideThe two tracks meet in one question: did the complete workflow become more useful, reliable, and economical—not simply faster to generate a draft?
For platform adoption, responsible AI principles, and workload architecture, see the AI engineering page. This page focuses on business choices and everyday use.
AI terms used in the course notes
- LLM / SLM
- Large / small language model: models that generate language from learned patterns, not guaranteed factual databases.
- Grounding / RAG
- Connecting an answer to reference information. Retrieval-augmented generation retrieves relevant sources before generating a response.
- Microsoft Graph / Work IQ
- Graph connects Microsoft 365 work data; Work IQ provides work context in supported Copilot experiences. Access still depends on permissions and configuration.
- DLP / least privilege
- Data loss prevention controls help restrict inappropriate sharing. Least privilege gives users and agents only the access needed for the task.
- KPI / ROI
- Key performance indicator / return on investment. Measure a useful business outcome and compare benefits with the full cost of achieving it.
- Tokens / credits
- Different usage units: tokens measure model input/output; credits measure consumption in applicable products. They are not interchangeable pricing units.
Strategy that can be operated
AB-731: from the business opportunity to solution selection, an operating model, and responsible scaling.
AB-731 connects five leadership decisions: value, solution choices, trust, pilots, and scaling. A continuing supplier-review case shows the evidence and ownership each decision requires.
AB-731: Read the lecture outline
Continue in the lecture: The main idea
Continue in the lecture: Generative AI foundations for leadersContinue in the lecture: IntroductionContinue in the lecture: What generative AI doesContinue in the lecture: Business value and solution familiesContinue in the lecture: Model choicesContinue in the lecture: Cost driversContinue in the lecture: Risks and opportunitiesContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Make AI useful and trustworthyContinue in the lecture: IntroductionContinue in the lecture: Prompt engineeringContinue in the lecture: Grounding and RAGContinue in the lecture: Data quality and securityContinue in the lecture: When machine learning fitsContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Copilot as a business capabilityContinue in the lecture: IntroductionContinue in the lecture: Copilot solution choicesContinue in the lecture: Apps, search, Researcher, and AnalystContinue in the lecture: Map processes to capabilitiesContinue in the lecture: Buy, extend, or buildContinue in the lecture: Responsible use in CopilotContinue in the lecture: Licensing and extensibility costsContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Foundry and custom business solutionsContinue in the lecture: IntroductionContinue in the lecture: The role of Microsoft FoundryContinue in the lecture: Map scenarios to AI servicesContinue in the lecture: Prebuilt tools, models, RAG, and tuningContinue in the lecture: Consumption and commitmentsContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Connect tools, platforms, and securityContinue in the lecture: IntroductionContinue in the lecture: AI capability vocabularyContinue in the lecture: Data science, ML, and deep learningContinue in the lecture: An outcome-led transformation approachContinue in the lecture: Copilot, agents, and work contextContinue in the lecture: Data and cloud foundationsContinue in the lecture: Govern, secure, and manage AIContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Build the business case and pilotContinue in the lecture: IntroductionContinue in the lecture: Business strategy and prioritizationContinue in the lecture: Technology and data strategyContinue in the lecture: Pilot design and evaluationContinue in the lecture: Organization and cultureContinue in the lecture: Governance and securityContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Make responsible AI operationalContinue in the lecture: IntroductionContinue in the lecture: Societal impact and sensitive usesContinue in the lecture: The six responsible AI principlesContinue in the lecture: Design governance responsibilitiesContinue in the lecture: Apply governance to build and buyContinue in the lecture: Microsoft's governance exampleContinue in the lecture: Sensitive-use review in practiceContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Scale with ownership and enablementContinue in the lecture: IntroductionContinue in the lecture: Move beyond isolated pilotsContinue in the lecture: Roles and responsibilitiesContinue in the lecture: Enable business usersContinue in the lecture: Enable domain expertsContinue in the lecture: Self-checkContinue in the lecture: Summary
Practical AI for business users
AB-730: prompting, content, analysis, collaboration, and bounded delegation in everyday work.
AB-730 connects five steps in everyday work: tasks, evidence, outputs, collaboration, and delegation. A continuing quarterly customer-review case shows how to brief AI, verify results, and make reliable handoffs.
Check the work account, license, tenant settings, region, and current product documentation before trying a feature. The reviewed material includes Copilot Memory, Prompt Lab, Work IQ, and Cowork; Code and Autopilot are described as preview-dependent. A feature appearing in training does not establish access in your environment.
AB-730: Read the lecture outline
Continue in the lecture: The main idea
Continue in the lecture: Prepare to use AI at workContinue in the lecture: IntroductionContinue in the lecture: How generated content differs from a recordContinue in the lecture: Responsible daily useContinue in the lecture: Chat, Cowork, Code, and AutopilotContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Know the Chat context and data boundaryContinue in the lecture: IntroductionContinue in the lecture: What Chat can accessContinue in the lecture: Access through the work accountContinue in the lecture: Navigate sources and conversationsContinue in the lecture: Enterprise data protectionContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Give AI a clear, reusable briefContinue in the lecture: IntroductionContinue in the lecture: Goal, context, sources, and expectationsContinue in the lecture: Iterate and verifyContinue in the lecture: Memory and custom instructionsContinue in the lecture: Prompt Lab and scheduled promptsContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Turn a useful chat into shared workContinue in the lecture: IntroductionContinue in the lecture: Summaries, drafts, files, and visualsContinue in the lecture: Copilot PagesContinue in the lecture: Purpose-built agentsContinue in the lecture: Practice: plan a client summitContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Draft, research, refine, and communicateContinue in the lecture: IntroductionContinue in the lecture: Word, PowerPoint, and OutlookContinue in the lecture: Researcher and source synthesisContinue in the lecture: Practice: a coordinated content packageContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Turn data into a defensible decisionContinue in the lecture: IntroductionContinue in the lecture: Analysis inside ExcelContinue in the lecture: Multi-file analysis with AnalystContinue in the lecture: Practice: sales and feedback analysisContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Make meetings produce reliable follow-throughContinue in the lecture: IntroductionContinue in the lecture: Before, during, and after the meetingContinue in the lecture: Chat, Pages, and NotebooksContinue in the lecture: Practice: meeting to project workspaceContinue in the lecture: Self-checkContinue in the lecture: SummaryContinue in the lecture: Delegate bounded work with CoworkContinue in the lecture: IntroductionContinue in the lecture: Chat, specialized agents, or CoworkContinue in the lecture: Consumption-based usageContinue in the lecture: Plan, progress, approvals, and resultsContinue in the lecture: Write a bounded assignmentContinue in the lecture: Evaluate realistic scenariosContinue in the lecture: Permissions and limitationsContinue in the lecture: Practice: review a delegated taskContinue in the lecture: Self-checkContinue in the lecture: Summary
Connect strategy to a real workflow
An illustrative pilot—not a claim of a completed project or a promised return.
Consider a monthly supplier review: approved service reports, invoices, and meeting notes become a decision brief. The goal is better-supported supplier decisions, not automatic negotiation, payment, or contract changes.
- Define the outcome
- Name the process owner and decision audience. Baseline preparation time, rework, factual errors, and overdue follow-ups using comparable past reviews.
- Bound the information
- Use an approved source folder and reporting period. Remove unnecessary personal data, check commercial confidentiality, and flag missing documents rather than inferring them.
- Choose the assistance
- Use Chat for a focused draft, Excel or Analyst for checked comparisons, and Pages or Notebooks for the working record. Delegate connected deliverables only where the enabled tool and approval settings are suitable.
- Keep human decisions
- Reconcile invoice totals, confirm reported service levels, inspect citations, and have the owner approve recommendations. Sending, sharing, booking, purchasing, and contract changes need separate authorization.
- Measure the whole process
- Track drafting plus checking and correction time, accepted-output rate, decision usefulness, and total cost. Keep any realized financial benefit separate from released staff capacity.
- Decide whether to scale
- Set target quality and cost thresholds before the trial. Expand only after repeated success with representative cases; pause for sensitive-data exposure, material errors, or rising review burden.
A bounded brief for the assistant
Prepare a supplier-review draft from the approved reports for the specified month. Compare reported service levels and charges, link each finding to its source, and list missing evidence separately. Produce a one-page decision brief and an action table with proposed owners. Do not send messages, change meetings, make purchases, or alter any source record. Ask me about unresolved scope before proceeding.
My acceptance question: can another reviewer trace the recommendation back to the same evidence and understand what the AI did not establish?
Scope and sources
Course references reviewed on .
The dedicated AB-731 and AB-730 pages each organize a lecture around five themes, connecting concepts across modules into a practical method. These are original business-oriented syntheses, not Microsoft's lesson text or an official course replacement. A complete index of the reviewed sources appears at the end.
Self-check links lead to Microsoft Learn; no assessment questions or answers are reproduced. The five linked exercise instructions were reviewed, not executed in a licensed tenant. Reading this page does not earn Microsoft training credit or establish course completion.
Course pages, product capabilities, and licensing can change independently. Where descriptions vary, verify the current product terms and your administrator's configuration; do not assume universal availability, unlimited consumption, automatic compliance, or guaranteed accuracy.
Official starting points: AB-731T00 (opens in a new tab) and AB-730T00 (opens in a new tab). Each course guide links to the official learning paths and every original module and unit.