AI Implementation for People Managers [2026]

Why AI Implementation Is a People Manager's Job Now

AI implementation is often treated as an IT or transformation responsibility, but the real impact of AI happens in the day-to-day work of teams. The people closest to that work are the ones who can see what needs to change, what should remain human, and where AI can genuinely help.

In this lecture, you’ll explore why AI implementation is fundamentally a people management responsibility. You’ll examine how managers make task-level decisions about AI, why simply introducing a tool does not create productivity, and how leadership judgment shapes whether AI improves work or creates new problems.

In This Lecture, You Will:

▸ Understand why AI implementation goes beyond strategy documents and technology decisions to affect the actual work teams perform every day.

▸ Recognize why people managers are uniquely positioned to identify where work gets stuck, which tasks matter, and where AI could change how work flows.

▸ Reframe the question from “How do I get my team using AI more?” to “How do I get this done, and should AI be doing part of it?”

▸ Evaluate AI task by task by considering what AI can do alongside your team's strengths, existing workload, and priorities.

▸ Identify why simply introducing an AI tool without changing workflows, approvals, or expectations can lead to low adoption, distrust, duplicated work, or unchanged bottlenecks.

▸ Consider how AI could free people from routine work such as summarising, drafting, and basic analysis while creating more time for judgment, customer conversations, collaboration, and thinking.

▸ Recognize the leadership decisions involved in shaping roles, responsibilities, human involvement, and team communication as AI changes work.

▸ Apply existing management skills such as judgment, knowledge of your team, clear communication, and decision-making to AI implementation.

Who This Lecture Is For:

• People Managers — Understand why AI implementation directly falls within your responsibility for how work gets done.

• Team Leaders — Identify where AI can change team tasks, workflows, and responsibilities.

• Department Managers — Evaluate AI implementation beyond simply introducing new tools.

• Managers Leading AI-Related Change — Understand the leadership judgment needed when AI changes how people work.

Why This Lecture Is Useful for People Managers

This lecture helps managers see AI implementation as a leadership responsibility rather than a technology rollout. By focusing on the work, the people doing it, and the decisions surrounding AI adoption, you can approach implementation with greater clarity and judgment.

What Good AI Implementation Looks Like (and What Bad Looks Like)

Buying an AI tool does not automatically improve how work gets done. Two teams can have the same work, the same starting point, and the same AI tool—and still end up with completely different results depending on what happens around the technology.

In this lecture, you’ll follow two insurance claims teams to see the difference between AI implementation that simply speeds up one part of an existing process and implementation that actually redesigns the work around AI. The comparison shows how process changes, role redesign, risk-based decisions, and clear communication can shape the outcome.

In This Lecture, You Will:

▸ Compare two teams that start with the same work, access to the same AI tool, and similar operating conditions but achieve very different outcomes.

▸ Identify how Company A creates a new bottleneck by making claim review faster without changing the rest of the process.

▸ Recognize how low trust can lead employees to redo AI-assisted work, while excessive trust can lead to mistakes through overly quick approval.

▸ Understand how Company B reorganises the claims process based on the complexity and risk of individual claims.

▸ Explore how lower-value claims can receive a lighter human review while complicated or high-value claims continue through a fuller review process.

▸ Examine how AI changes people's roles by reducing routine first-pass work and giving team leads more time for complex judgment and coaching.

▸ Recognize why communicating what is changing, why it is changing, and what it means for employees is an important part of implementation.

▸ Distinguish between simply adding AI to an existing workflow and rethinking the process, roles, and team experience around the technology.

▸ Reflect on how your own team could unintentionally fall into the Company A pattern—and what deliberate decisions could move implementation toward the Company B approach.

Who This Lecture Is For:

• People Managers — See how implementation decisions affect workflows, roles, workload, and team trust.

• Team Leaders — Understand how AI can change where work and review responsibilities sit within a team.

• Department Managers — Learn to look beyond the AI tool itself and examine the process surrounding it.

• Managers Leading AI Implementation — Recognize the practical differences between simply deploying AI and redesigning work around it.

Why This Lecture Is Useful for People Managers

This lecture gives you a practical picture of what successful AI implementation looks like. By comparing Company A and Company B, you can see why the quality of implementation depends less on the AI tool itself and more on how managers rethink processes, match human review to risk, reshape roles, and communicate change.

Key AI Glossary — The Words You'll Keep Hearing

AI conversations can quickly become confusing when terms like AI, machine learning, generative AI, agents, and automation are used as though they all mean the same thing. For people managers, understanding the vocabulary is less about becoming technical and more about being able to confidently discuss AI with your team and make sense of what different systems actually do.

In this lecture, you’ll build a practical reference for the AI terms you’ll keep encountering throughout the course. You’ll distinguish different types of AI, understand the difference between prediction and human judgment, and explore concepts such as agents, hallucinations, and Retrieval-Augmented Generation (RAG).

In This Lecture, You Will:

▸ Define artificial intelligence as a broad category of systems that can perform tasks that normally require human thinking.

▸ Distinguish predictive AI, generative AI, automation or agentic AI, and decision-support AI based on how they work and how they affect managerial oversight.

▸ Understand how predictive AI uses past data to estimate what is likely to happen next.

▸ Recognize how generative AI creates new content such as text, images, code, and summaries.

▸ Explore how agentic AI can carry out multiple steps of a task rather than simply providing a recommendation or answer.

▸ Understand decision-support AI and why recommendations, scores, or options still require a person to make the final decision.

▸ Identify what an AI agent does and why greater autonomy makes it important to know how to intervene when something goes wrong.

▸ Recognize AI hallucinations and why a confident, polished response is not necessarily a correct one.

▸ Differentiate prediction from judgment by examining where AI-generated probabilities end and a manager's human decision-making begins.

▸ Understand the basic idea behind Retrieval-Augmented Generation (RAG) and how it can ground AI responses in an organisation's own policies, reports, and internal knowledge.

Who This Lecture Is For:

• People Managers — Build the vocabulary needed to confidently discuss AI systems and their impact on your team.

• Team Leaders — Understand the different types of AI your team may encounter and the level of oversight they may require.

• Department Managers — Distinguish between AI that predicts, generates, acts, or supports decisions.

• Managers Leading AI Adoption — Develop a practical understanding of common AI terminology used in implementation discussions.

Why This Lecture Is Useful for People Managers

This lecture gives you a practical AI vocabulary without requiring technical knowledge. Understanding these terms helps you recognise what an AI system is actually doing, where human judgment remains necessary, and what different types of AI may mean for your team's work and accountability.

The State of Generative AI

Generative AI has moved far beyond simply answering prompts. Today’s tools can research information, reason through complex tasks, interact with websites, create business deliverables, and even build working applications from natural-language instructions.

In this lecture, you’ll see two practical demonstrations that show how far these capabilities have developed. You’ll explore ChatGPT Agent Mode through a recruitment planning task and Google AI Studio Build through a working CV screening application.

In This Lecture, You Will:

▸ Explore how generative AI has evolved from simple text responses to completing complex, multi-step professional tasks.

▸ Observe ChatGPT Agent Mode handling a recruitment brief involving labor market research, sourcing channels, budget estimation, and recruitment planning.

▸ See how the agent uses reasoning, web research, and a virtual environment to work through the recruitment task.

▸ Review the PowerPoint recruitment strategy generated from the task, including hiring objectives, market and salary benchmarks, sourcing strategy, budget, timeline, and next steps.

▸ Examine the Excel hiring budget, including estimated cost per hire, individual cost components, total budget, and documented assumptions.

▸ Explore Google AI Studio Build through an HR use case for screening candidate CVs.

▸ See Gemini create a working application that accepts a job title, job description, and candidate CVs in PDF format.

▸ Examine the application's CV analysis, including a match score, candidate summary, matching qualifications, and missing requirements.

▸ Observe how Gemini can fix a coding error through a confirmation and continue working on the application.

▸ See how the application can be updated through follow-up prompts, including adding the ability to upload multiple CVs.

▸ Recognize how generative AI is moving beyond answering questions toward researching, reasoning, building, and producing practical professional outputs.

Tools Covered:

• ChatGPT Agent Mode

• Google AI Studio Build

Who This Lecture Is For:

• People Managers — See how current generative AI capabilities can support complex professional work.

• HR and Talent Leaders — Explore practical examples of AI applied to recruitment planning and CV screening.

• Team Leaders — Understand how AI can handle multi-step tasks and support practical application development.

• Managers Exploring AI Adoption — Gain a practical view of what current generative AI tools can accomplish.

Why This Lecture Is Useful for People Managers

This lecture provides a practical view of what generative AI can currently do through two real demonstrations. By seeing AI research, business deliverables, and application development in action, you can better understand the capabilities that are becoming available to professionals and teams.

RAG: What It Is and Why It Matters

AI can give fast and convincing answers, but what happens when a People Manager needs an answer based on company-specific policies, documents, or internal knowledge? Retrieval-Augmented Generation (RAG) helps address this by connecting AI responses to trusted organisational information.

In this lecture, you’ll learn how RAG retrieves relevant information from a knowledge base and uses it to generate more grounded responses. You’ll also understand the key concepts behind chunking, embeddings, vector databases, and semantic retrieval.

In This Lecture, You Will:

▸ Understand what Retrieval-Augmented Generation (RAG) is and why it matters for organisational AI use.

▸ Explore how a knowledge base can use documents such as PDFs, text files, and CSVs to provide trusted information.

▸ Understand why sending an entire document to an AI model can increase cost, exceed context limits, and introduce irrelevant information.

▸ Learn how documents are divided into meaningful chunks so that relevant information can be retrieved.

▸ Understand how embeddings represent the meaning of text and enable similarity-based retrieval.

▸ Explore how vector databases store and retrieve information based on semantic similarity rather than exact keyword matches.

▸ Follow the basic flow of retrieving relevant information and using it to generate a grounded AI response.

Who This Lecture Is For:

• People Managers — Understand how AI can use trusted company information when answering workplace-related questions.

• AI & Product Managers — Build a foundation for understanding how organisational knowledge can be connected to AI systems.

• Innovation & Technology Leaders — Understand the core concepts behind knowledge-grounded AI.

• Professionals Working with AI Systems — Learn the terminology and workflow behind RAG-based systems.

Why This Lecture Is Useful for People Managers

People Managers often work with company policies, processes, and internal information. Understanding RAG helps them recognise how AI systems can provide answers grounded in trusted organisational knowledge rather than relying only on general information.

RAG in Action

RAG makes more sense when you see it working with real data. In this lecture, you’ll build a simple RAG system using Python and see how a design system document can be turned into a searchable knowledge source for AI.

You’ll follow the complete process step by step—from uploading the document and creating chunks to generating embeddings, storing them in a vector store, retrieving relevant information, and generating a grounded answer through ChatGPT.

In This Lecture, You Will:

▸ Build a simple RAG system using Python and follow the implementation step by step.

▸ Upload a product design system document and make its content available to the RAG system.

▸ Split the document into smaller chunks with overlap to preserve important context.

▸ Convert document chunks and user questions into embeddings for similarity-based retrieval.

▸ Store embeddings in a Chroma vector store to make the design system searchable.

▸ Retrieve the most relevant document chunks based on a user's question.

▸ Connect retrieved information to a prompt template that instructs the AI to answer using the provided context.

▸ Generate a grounded answer through ChatGPT based on the actual design system documentation.

▸ See how RAG can help teams search documents such as design guidelines, product requirements, brand guidelines, and accessibility standards.

Tools Covered:

• Python

• Google Colab

• OpenAI API

• LangChain

• Chroma

• ChatGPT

Features & Techniques Covered:

• Document Upload

• Document Chunking

• Embeddings

• Vector Store

• Similarity-Based Retrieval

• Retriever Pipeline

• Prompt Template

• Grounded AI Responses

Who This Lecture Is For:

• People Managers — Understand how organisational documents can be connected to AI to make internal information easier to access.

• AI & Product Managers — Learn how RAG can turn documentation into a searchable AI knowledge source.

• Innovation & Technology Leaders — Understand the practical workflow behind a RAG implementation.

• Professionals Working with AI Systems — Follow a complete example of building and testing a RAG system.

Why This Lecture Is Useful for People Managers

People Managers can see how RAG can make important internal documentation easier for teams to access and query in plain language. The practical demonstration helps them understand how trusted organisational information can be connected to AI instead of relying only on general AI knowledge.

Prediction Is Not Judgment — The Leadership Decision Test

AI can analyse huge amounts of data and make highly confident recommendations. But when a decision involves uncertainty, context, relationships, values, and consequences, should a leader simply follow the prediction?

In this lecture, you’ll explore the critical difference between AI prediction and human judgment through realistic leadership scenarios, including AI-assisted hiring and customer service staffing decisions. You’ll learn how leaders can evaluate AI recommendations within the broader context before making the final decision.

In This Lecture, You Will:

▸ Distinguish between AI prediction and human judgment.

▸ Understand how AI uses historical data and patterns to predict likely outcomes.

▸ Recognize why leadership decisions often require context, relationships, values, trade-offs, and incomplete information.

▸ Evaluate an AI hiring recommendation when your own assessment raises concerns.

▸ Examine how changing circumstances can affect the usefulness of an AI recommendation based on historical data.

▸ Apply the principle of combining AI-generated insights with human judgment rather than blindly trusting or dismissing AI outputs.

▸ Reflect on your own significant decisions and identify where data alone was not enough.

Who This Lecture Is For:

• People Managers — Learn how to evaluate AI recommendations while keeping human judgment at the centre of important decisions.

• Team Leaders — Understand when AI predictions can inform a decision and when broader context needs to be considered.

• HR & Talent Leaders — Explore the difference between data-driven recommendations and leadership judgment in people decisions.

• Business Leaders — Develop a balanced approach to using AI-generated insights in uncertain and high-impact decisions.

Why This Lecture Is Useful for People Managers

People Managers increasingly encounter AI-generated recommendations about people, performance, staffing, and other workplace decisions. This lecture helps them understand that a strong prediction is not automatically the right decision—and that their responsibility is to evaluate AI recommendations within the real-world context before deciding.

The 5-Minute AI Audit

The question for leaders is no longer simply “Should we use AI?” It is “Where exactly should we use AI?” This lecture introduces a simple five-minute audit to help People Managers identify which team activities AI should own, assist with, or stay out of.

You’ll evaluate your team’s recurring activities using three practical questions: whether the work is repetitive, whether there is usually a correct answer, and whether experienced managers would generally reach the same conclusion.

In This Lecture, You Will:

▸ Identify 5–10 recurring activities your team performs regularly.

▸ Assess whether each activity is repetitive and follows a consistent process.

▸ Evaluate whether an activity has a clear and objectively verifiable outcome.

▸ Consider whether experienced managers would generally reach the same conclusion.

▸ Classify activities as strong candidates for AI ownership, AI assistance, or human-led work.

▸ Apply the audit to examples such as reviewing notes, customer FAQ responses, business reviews, and succession planning.

▸ Recognize why activities involving greater disagreement, complexity, and competing values require more human judgment.

▸ Identify one activity to automate, one where AI can assist, and one that should remain firmly human-led.

Who This Lecture Is For:

• People Managers — Identify where AI can meaningfully support your team's everyday work.

• Team Leaders — Evaluate activities based on the amount of repetition and human judgment they require.

• HR & Talent Leaders — Distinguish between activities suitable for AI support and decisions that require leadership judgment.

• Business Leaders — Use a work-first approach to determine where AI and automation are appropriate.

Why This Lecture Is Useful for People Managers

The audit gives People Managers a practical way to examine their team's actual work before deciding where AI belongs. It helps them separate repetitive activities that AI can own, tasks where AI can assist, and work where human judgment should remain central.

The Automation Trap — Capability vs. Consequence

AI may be capable of performing a task perfectly, but that does not automatically mean it should be given full control. For People Managers, the more important question is what could happen if the AI gets the decision wrong.

In this lecture, you’ll explore the difference between AI capability and leadership responsibility and learn how to evaluate automation based on risk, impact, and accountability.

In This Lecture, You Will:

▸ Distinguish between asking whether AI can perform an activity and evaluating whether it should.

▸ Recognize how delegating work to AI also means delegating a certain level of risk.

▸ Evaluate the potential consequences of AI errors across different types of activities.

▸ Assess the cost of being wrong, who is affected by the outcome, and who remains responsible.

▸ Compare low-risk activities such as meeting summaries with higher-risk decisions such as employee promotions and corporate strategy.

▸ Determine when AI can automate a task, when human review is needed, and when humans should remain responsible for the decision.

▸ Apply the risk assessment questions to an activity previously identified as an automation candidate.

▸ Recognize why effective AI adoption depends on better judgment about where AI should and should not be used.

Who This Lecture Is For:

• People Managers — Evaluate the consequences of delegating team activities and decisions to AI.

• Team Leaders — Understand when automation requires human oversight based on risk and impact.

• HR & Talent Leaders — Consider the consequences of using AI in decisions that affect employees and careers.

• Business Leaders — Make more informed decisions about automation by considering risk, impact, and accountability.

Why This Lecture Is Useful for People Managers

People Managers are responsible for the outcomes of decisions made within their teams, even when AI is involved. This lecture helps them look beyond what AI is capable of and consider the cost of errors, who could be affected, and where human accountability must remain.

Case Study: Air Canada's Chatbot

AI can make customer service faster and more efficient—but when an AI system gives incorrect information, who is responsible for the outcome? The Air Canada chatbot case provides a real-world example of why delegating a task does not mean delegating accountability.

In this lecture, you’ll examine what happened when Air Canada’s chatbot provided incorrect information about bereavement fares and explore the leadership lessons around AI delegation, risk, human oversight, and accountability.

In This Lecture, You Will:

▸ Examine the Air Canada chatbot case and the customer service situation that led to legal review.

▸ Understand why the chatbot's incorrect response created consequences for the organization.

▸ Recognize that customers interact with the organization through its AI system, not with an independent entity.

▸ Distinguish between delegating activities and retaining accountability for outcomes.

▸ Evaluate the cost of an AI error, who could be affected, how easily the mistake can be detected, and who remains accountable.

▸ Determine why different AI activities require different levels of human oversight.

▸ Recognize why AI capability alone should not determine whether AI should own an outcome.

Who This Lecture Is For:

• People Managers — Understand the accountability that remains with leaders when AI performs work on behalf of the organization.

• Team Leaders — Evaluate when AI tasks require human review and oversight.

• HR & Talent Leaders — Apply the principles of risk, accountability, and oversight when considering AI delegation.

• Business Leaders — Understand why efficiency and accountability must be considered separately when using AI.

Why This Lecture Is Useful for People Managers

People Managers may delegate tasks to AI, but they remain responsible for how those tasks affect employees, customers, and the organization. This case helps them understand why AI delegation must be matched with appropriate oversight and clear human accountability.

The AI Delegation Framework — Four Levels

AI should not automatically take over every task, and humans should not automatically handle everything themselves. The real leadership challenge is deciding who should do what, and how much responsibility AI should have.

In this lecture, you’ll learn a practical four-level delegation framework that helps leaders deliberately decide where different activities belong—from human-led work to full AI automation.

In This Lecture, You Will:

▸ Understand the four levels of AI delegation and how responsibility shifts between humans and AI.

▸ Identify when activities should remain primarily human-led because judgment, ethics, relationships, or accountability are critical.

▸ Apply Level 2, where AI assists with preparation and analysis while humans remain responsible for decisions and outcomes.

▸ Recognize Level 3 workflows where AI executes the task but humans review and approve the output before action is taken.

▸ Evaluate Level 4 automation for repetitive, predictable activities where occasional errors are manageable and correctable.

▸ Distinguish optimal delegation from simply trying to maximize automation.

▸ Assign three activities from your own work to the appropriate delegation level.

▸ Consider how deliberate delegation can help leaders redesign jobs, restructure teams, and rebuild processes around higher-value human work.

The Four Levels Covered:

• Level 1 — Human Performs: Humans remain the primary decision makers when judgment, ethics, relationships, or accountability are central.

• Level 2 — AI Assists: AI prepares information, analysis, drafts, or options while humans make the decisions.

• Level 3 — AI Executes with Human Review: AI performs the task, but a human reviews and approves the output before it is finalized or acted upon.

• Level 4 — AI Automates: AI performs the activity independently, with humans monitoring performance and handling exceptions.

Who This Lecture Is For:

• People Managers — deciding how responsibilities should be divided between teams and AI.

• Team Leaders — evaluating which team activities can be assisted, reviewed, or automated.

• HR & Talent Leaders — considering where human judgment should remain central in people-related work.

• Business Leaders — making deliberate decisions about AI delegation and automation.

Why This Lecture Is Useful for People Managers

People Managers need more than an understanding of what AI can do—they need a way to decide how much responsibility AI should actually have. This framework gives them a practical way to classify team activities, preserve human judgment where it matters, and identify appropriate opportunities for AI assistance or automation.

Why AI Often Fails — The Leadership Problem

When AI produces a generic, incomplete, or unhelpful result, it is easy to blame the technology. But often, the real issue is that the objective, context, or expectations were never clearly defined.

In this lecture, you’ll learn how clearer leadership translates into better AI outputs and a practical four-question framework for defining significant AI tasks before handing them over.

In This Lecture, You Will:

▸ Recognize why unclear instructions, vague objectives, and undefined success criteria can lead to poor AI outputs.

▸ Understand how clearer leadership can improve the usefulness of AI-generated work without changing the underlying technology.

▸ Apply the same principles used when delegating work effectively to team members when working with AI.

▸ Identify how AI can expose existing weaknesses in management practices, particularly unclear objectives and expectations.

▸ Use the four-question leadership prompt engineering checklist:

• Define the objective — What business outcome are you trying to achieve?

• Identify the audience — Who will use the output and what do they need?

• Provide context — What information, constraints, format, tone, scope, and exclusions matter?

• Define success — What would a genuinely useful output look like?

▸ Practice applying the four questions to an activity you have identified as suitable for AI.

▸ Recognize how clarifying expectations can improve both AI-generated work and your own thinking.

▸ Understand why effective AI adoption is not only a technology challenge, but also a management challenge.

Who This Lecture Is For:

• People Managers — improving how they define and delegate work to AI.

• Team Leaders — giving clearer objectives, context, and expectations for AI-supported tasks.

• HR & Talent Leaders — using AI more effectively while maintaining clear outcomes and expectations.

• Business Leaders — improving the value they extract from AI through clearer leadership.

Why This Lecture Is Useful for People Managers

People Managers regularly delegate work, and the quality of the result depends heavily on how clearly the work is defined. This lecture helps managers apply that same leadership discipline when working with AI—clarifying the objective, audience, context, and definition of success before asking AI to perform a task.

Case Study: Uber's Surge Pricing, Sydney Siege

What happens when an AI system follows its instructions perfectly—but those instructions fail to account for a critical situation? The Uber surge pricing case shows how a system can function exactly as designed and still produce a deeply problematic outcome.

In this lecture, you’ll examine the 2014 Sydney siege case and connect it to the leadership responsibility of defining objectives, setting constraints, anticipating exceptions, and understanding what AI cannot see.

In This Lecture, You Will:

▸ Examine Uber’s surge pricing response during the 2014 Sydney siege and how the algorithm responded to a sudden increase in demand.

▸ Understand why the algorithm was functioning as designed rather than experiencing a technical failure.

▸ Identify how an incomplete objective can produce unintended outcomes when important constraints are missing.

▸ Recognize why AI cannot automatically understand changing circumstances, emergencies, or contextual differences unless these are accounted for in its instructions and design.

▸ Connect the case to the four leadership questions of objective, audience, context, and success.

▸ Evaluate the importance of anticipating edge cases, contextual exceptions, and situations that an AI system may not be able to recognize.

▸ Understand why handing execution to AI does not mean handing over leadership responsibility or accountability.

▸ Recognize the importance of defining objectives clearly and building appropriate human guardrails before AI operates independently.

Who This Lecture Is For:

• People Managers — understanding the leadership responsibility behind AI-driven decisions and outcomes.

• Team Leaders — learning how incomplete instructions and missing constraints can affect AI-supported work.

• Business Leaders — considering objectives, context, exceptions, and accountability when delegating execution to AI.

• AI & Innovation Leaders — examining how AI systems can produce unintended outcomes even when functioning as designed.

Why This Lecture Is Useful for People Managers

People Managers may delegate execution to AI, but they remain responsible for defining what the system should achieve and the boundaries within which it should operate. This case demonstrates why managers must think beyond the normal workflow and anticipate situations where context changes, exceptions arise, or a perfectly functioning system could still produce the wrong outcome.

Don't Trust the First Answer

AI can produce answers that look polished, professional, and convincing—but a confident answer is not necessarily a correct or useful one. The responsibility for evaluating what AI produces still belongs to the human.

In this lecture, you’ll learn a practical five-question framework for evaluating AI outputs before acting on them, with a focus on accuracy, completeness, relevance, reasonableness, and accountability.

In This Lecture, You Will:

▸ Recognize why a confident and well-structured AI response should not automatically be treated as correct.

▸ Understand the difference between AI generating an output that looks right and a human determining whether it is actually right.

▸ Evaluate AI-generated work for accuracy, completeness, relevance, and practical fit.

▸ Analyze an AI recommendation to reduce customer support staffing and identify what additional evaluation may be required before acting.

▸ Practice applying the five-question framework to an AI output or use case from your own work.

▸ Recognize how consistent evaluation changes the role of AI from an answer machine into a thinking partner that supports better decisions.

▸ Understand why effective leaders evaluate AI based on whether it helps them make better decisions, not simply whether it produces answers quickly.

Who This Lecture Is For:

• People Managers — evaluating AI-generated recommendations and outputs before using them in consequential decisions.

• Team Leaders — developing a consistent process for reviewing AI-supported work.

• HR & Talent Leaders — applying human judgment when AI outputs influence people-related decisions.

• Business Leaders — ensuring AI-generated recommendations are accurate, relevant, reasonable, and defensible.

Why This Lecture Is Useful for People Managers

People Managers remain accountable for decisions even when AI provides the recommendation or output. This lecture gives managers a practical way to slow down before acting, identify gaps in AI-generated work, and apply their own judgment to determine whether an output is accurate, relevant, reasonable, and safe to act on.

Case Study: The Lawyer Who Cited Cases That Never Existed

AI can produce information that looks authoritative, professional, and convincing—even when that information is completely false. This case study shows what can happen when professionals trust polished AI output without verifying it.

In this lecture, you’ll examine the 2023 case involving lawyers who used ChatGPT for legal research, submitted fabricated cases in a court filing, and faced serious consequences. The case demonstrates why verification and human judgment remain essential when using AI for professional work.

In This Lecture, You Will:

▸ Examine how lawyers used ChatGPT to identify legal cases that could support their arguments.

▸ Recognize how fabricated legal cases were presented by AI as legitimate precedents.

▸ Understand how professional-looking language, credible-sounding citations, and confident output can create a false sense of accuracy.

▸ Identify the larger professional failure of accepting AI-generated information without verification.

▸ Distinguish confidence from accuracy, fluency from verification, and polished output from trustworthy information.

▸ Apply the evaluation questions from the previous lecture by considering whether AI-generated information is accurate, complete, relevant, reasonable, and something you are willing to be accountable for.

▸ Recognize how a simple verification step could have prevented the problem in this case.

▸ Understand why AI-generated outputs should be treated as inputs into human decision-making rather than final answers.

▸ Recognize why evaluation and discernment are critical leadership skills when using AI in professional work.

Who This Lecture Is For:

• People Managers — developing stronger judgment when reviewing AI-generated information before it is used.

• Team Leaders — ensuring AI-supported work is evaluated rather than accepted at face value.

• Business Leaders — understanding the professional risks of acting on unverified AI outputs.

• Professionals Working with AI — learning why polished AI-generated information still requires human validation.

Why This Lecture Is Useful for People Managers

People Managers may rely on AI-generated information for reports, analyses, proposals, forecasts, or recommendations. This case reinforces that managers remain responsible for determining whether AI-generated information deserves confidence before it is used or acted upon.

What AI Can't See — Blind Spots

An AI recommendation can be accurate, logical, and supported by solid data—and still lead to the wrong decision. The reason is simple: AI sees patterns, while leaders must see context.

In this lecture, you’ll explore four areas AI can struggle to fully account for: context, stakeholders, ethics, and long-term consequences. You’ll learn how these blind spots can change the meaning of an otherwise reasonable AI recommendation and where human judgment becomes essential.

In This Lecture, You Will:

▸ Understand why an AI recommendation can be data-supported and still result in the wrong decision.

▸ Explore how context can change the judgment behind an AI recommendation, using an employee performance example.

▸ Identify four AI blind spots: context, stakeholders, ethics, and long-term consequences.

▸ Recognize why AI only knows the context it receives and may miss organizational history, team dynamics, customer relationships, or recent events.

▸ Evaluate how stakeholders and competing interests can be overlooked when AI optimizes for a specific task or objective.

▸ Examine why effective AI use requires human judgment about fairness, appropriateness, responsibility, and potential unintended consequences.

▸ Consider how AI recommendations based on historical patterns can overlook long-term effects that are not visible in current data.

▸ Apply four questions to decisions involving hiring, budgeting, project prioritization, customer management, and team leadership.

▸ Recognize how ambiguity, competing trade-offs, relationships, ethics, and future uncertainty create areas where human judgment remains essential.

▸ Apply the question, “What might the AI not be seeing?” when evaluating an AI recommendation.

Who This Lecture Is For:

• People Managers — making decisions where employee context, relationships, ethics, and long-term consequences matter.

• Team Leaders — recognizing when AI recommendations may not capture the full situation surrounding a decision.

• HR & Talent Leaders — evaluating AI-supported decisions involving employees, stakeholders, and workplace outcomes.

• Business Leaders — balancing data-driven recommendations with broader organizational considerations.

Why This Lecture Is Useful for People Managers

People Managers make decisions that cannot always be reduced to measurable data. This lecture helps managers recognize what AI may miss—context, affected stakeholders, ethical considerations, and long-term consequences—so they can identify where their own judgment adds value before accepting an AI recommendation.

Hallucination & Bias in Generative AI

When generative AI becomes part of a product design process, its outputs can directly shape the experiences of real users. But biased outputs can exclude important user groups, while hallucinated information can lead teams to make decisions based on facts that were never verified.

In this lecture, you’ll explore how bias and hallucination can affect AI-assisted product design and examine four practical strategies for reducing these risks: inclusive inputs, human review, ongoing monitoring, and Retrieval-Augmented Generation (RAG).

In This Lecture, You Will:

▸ Understand how bias and hallucination can affect AI-assisted product design and ultimately shape real user experiences.

▸ Explore how biased training data can cause AI to reproduce narrow assumptions about users, demographics, abilities, geographies, and use contexts.

▸ Recognize how biased personas, onboarding recommendations, and UX patterns can influence design decisions in ways that exclude parts of the actual user base.

▸ Identify hallucination as the generation of factually incorrect or fabricated information that is presented with confidence.

▸ Examine how fabricated benchmarks or incorrect accessibility guidance can lead product teams to make decisions based on unverified information.

▸ Understand why AI-generated outputs should not be treated as finished deliverables, particularly when they influence high-stakes design decisions.

▸ Apply inclusive input design by deliberately requesting diverse user representations and testing AI-generated outputs across different user groups.

▸ Build human review into AI-assisted design workflows so that experienced designers or researchers validate important outputs before they influence the direction of the work.

▸ Recognize why monitoring AI outputs needs to continue after launch, including collecting and reviewing user feedback about outputs that are wrong, biased, or confusing.

▸ Understand how RAG can ground AI responses in verified design knowledge such as design system documentation, validated research findings, and established UX guidelines.

▸ Distinguish between AI that generates plausible answers from general training data and AI that retrieves verified information from curated sources.

▸ Recognize the role of product designers in connecting AI capability with user trust and ensuring AI-powered features are accurate, fair, and designed with user needs in mind.

Who This Lecture Is For:

• Product Managers — ensuring AI-powered product features are designed with accuracy, fairness, human oversight, and user trust in mind.

• Product Designers — identifying and reducing bias and hallucination in AI-assisted personas, UX content, design recommendations, and product experiences.

• UX Researchers — validating AI-generated outputs and considering diverse user groups when evaluating AI-assisted design work.

• Design Leaders — establishing review, monitoring, and responsible AI practices within product design workflows.

Why This Lecture Is Useful for Product Managers

Product Managers help shape AI-powered features that directly affect users. This lecture shows why responsible AI is not only a technical concern, but also a product and co-design responsibility—requiring inclusive inputs, human oversight, continuous monitoring, and verified knowledge to maintain user trust.

Case Study: Zillow Offers

What happens when an AI-driven business model works well—until the world changes faster than the data can explain? The Zillow Offers case shows how a company can rely on sophisticated predictive models, yet still face major losses when historical patterns stop reflecting reality.

In this lecture, you’ll examine how Zillow used algorithms and predictive models to estimate home values and make purchasing decisions, how the COVID-19 pandemic fundamentally changed housing-market conditions, and why human judgment was needed to recognize that the model’s historical assumptions were no longer reliable.

In This Lecture, You Will:

▸ Examine how Zillow Offers used algorithms and predictive models to estimate home values, purchase properties, improve them, and resell them for profit.

▸ Understand why Zillow's large amount of historical housing data and predictable pricing patterns made the AI-driven business model appear compelling.

▸ Explore how the COVID-19 pandemic changed remote work, migration, housing demand, supply, interest rates, and renovation costs in ways historical data had not anticipated.

▸ Recognize how Zillow's models continued relying on historical patterns even as the housing market fundamentally changed.

▸ Understand why the algorithm was not necessarily “broken,” but was operating in conditions that were fundamentally different from those represented in its training data.

▸ Learn about the concept of distributional shift—when conditions become so different from the training data that historical patterns are no longer reliable guides to the future.

▸ Identify why an AI system can analyze historical prices, market trends, comparable properties, and transaction data without recognizing that the broader environment has fundamentally changed.

▸ Connect the Zillow case to the importance of context and the leadership responsibility to recognize changes surrounding the data.

▸ Apply the lesson to AI-driven forecasts and recommendations involving customer demand, hiring, investment, market expansion, and revenue.

▸ Evaluate whether the conditions that generated historical data are still the conditions in which the organization is operating.

▸ Recognize why historical accuracy does not guarantee future success and why leaders must be prepared to question confident AI outputs when reality is changing.

Who This Lecture Is For:

• People Managers — recognizing when changing circumstances may make AI-driven recommendations less reliable.

• Business Leaders — understanding the limitations of historical data when markets, behaviors, or operating conditions change.

• Team Leaders — questioning whether the assumptions behind AI-driven forecasts still hold.

• Professionals Using AI for Decision-Making — learning to combine AI-generated analysis with broader context and human judgment.

Why This Lecture Is Useful for People Managers

People Managers may use AI-driven forecasts and recommendations to support decisions about demand, hiring, budgets, customers, or business priorities. The Zillow case reinforces that managers must look beyond what historical data says and recognize when changing conditions make those patterns unreliable.

The Feedback Loop Advantage

When AI gives you a poor answer, the most valuable response may not be to stop using it. High-performing professionals treat AI as a system that can improve through interaction, feedback, and better inputs rather than expecting a perfect result from the first prompt.

In this lecture, you’ll learn how to build a feedback loop around AI—from defining the objective and evaluating the initial output to providing specific feedback and repeating the process until the result becomes genuinely useful. You’ll also see how organizations can apply the same approach to AI systems such as customer service chatbots.

In This Lecture, You Will:

▸ Understand why high-performing professionals treat AI as a system and feedback loop rather than simply a question-and-answer tool.

▸ Recognize why AI outputs should be treated as starting points rather than expecting perfection from the first attempt.

▸ Explore how specific feedback, additional context, clearer requirements, and precise instructions can progressively improve an AI-generated result.

▸ Examine a customer communication strategy example where repeated feedback improves the output across multiple iterations.

▸ Understand why the AI itself does not necessarily become smarter during the interaction—the human improves the system by providing better input, clearer context, and more precise feedback.

▸ Apply the five-step feedback loop: define the objective, generate an initial output, evaluate the result, provide specific feedback, and improve and repeat.

▸ Evaluate AI outputs for accuracy, completeness, relevance, missing information, and areas that do not fit.

▸ Recognize how feedback loops can be applied beyond individual productivity to organizational AI systems.

▸ Examine how a customer service chatbot can improve by analyzing unanswered questions, missing knowledge, escalation triggers, and gaps between customer needs and what the system was designed to handle.

▸ Understand why improving AI systems requires human judgment to analyze results, interpret feedback, decide what to change, and evaluate whether those changes worked.

▸ Apply feedback-loop thinking to an AI use case by identifying missing context, useful feedback, lessons learned, and opportunities for improvement over time.

▸ Recognize why human responsibilities such as defining objectives, evaluating outputs, identifying blind spots, providing feedback, and deciding when a system needs to change become more important as AI becomes more capable.

Who This Lecture Is For:

• People Managers — building deliberate processes for improving AI-supported work through feedback and iteration.

• Team Leaders — helping teams move beyond one-shot AI interactions toward continuous improvement.

• Business Leaders — understanding how feedback systems can improve organizational AI use cases over time.

• Professionals Using AI — learning how to get more valuable results by evaluating, refining, and repeating AI interactions.

Why This Lecture Is Useful for People Managers

People Managers are responsible not only for using AI, but for creating processes that produce better outcomes over time. This lecture shows how structured feedback keeps human judgment at the center of AI use—helping managers define objectives, evaluate results, identify gaps, provide better inputs, and continuously improve the way AI supports their teams.

Case Study: Netflix's Recommendation Engine

AI systems do not have to be perfect from the beginning to become highly effective. Netflix’s recommendation engine demonstrates how continuous learning from user behavior can make AI outputs more relevant and personalized over time.

In this lecture, you’ll examine how Netflix uses feedback from user behavior and connect that idea to AI use in the workplace. You’ll also see how managers and organizations can improve AI-supported work by creating stronger feedback loops around the technology.

In This Lecture, You Will:

▸ Examine how Netflix's recommendations evolve from relatively generic suggestions to more relevant and personalized recommendations as the system learns from user behavior.

▸ Understand how actions such as watching, stopping, or repeatedly viewing particular types of content become feedback signals.

▸ Recognize why Netflix's success did not depend on having a perfect recommendation system from day one, but on continuously learning from user behavior.

▸ Compare two managers using AI to create customer presentations and see how different levels of feedback can produce different results with the same technology.

▸ Explore how specific feedback—such as focusing on customer retention, adding implementation risks, making recommendations more strategic, and using executive-level language—can improve AI-generated work.

▸ Recognize why productivity gains can come from the quality of the feedback loop rather than simply from the AI technology itself.

▸ Examine how organizations can treat customer service chatbot failures as learning opportunities by identifying unanswered questions, customer frustration, escalations, and missing information.

▸ Understand how feedback from chatbot interactions can be used to improve the system, customer experience, and organizational performance.

▸ Recognize why leaders create value not only by using AI, but also by identifying what works, what does not work, and what needs refinement.

▸ Understand why human judgment remains critical in evaluating AI outputs, providing feedback, and deciding what should change.

▸ Apply the central feedback-loop principle: AI generates outputs, humans evaluate them, humans provide feedback, and the overall system improves through that feedback.

Who This Lecture Is For:

• People Managers — improving AI-supported work by using deliberate feedback and continuous refinement.

• Team Leaders — helping teams build stronger feedback loops around AI tools and workflows.

• Business Leaders — understanding how continuous learning can improve organizational AI systems.

• Professionals Using AI — learning why better feedback processes can produce better outcomes from the same AI technology.

Why This Lecture Is Useful for People Managers

People Managers can improve the value their teams get from AI by focusing not only on the technology, but also on how the team evaluates and improves its outputs. The Netflix case demonstrates why continuous feedback, human evaluation, and deliberate refinement are central to creating better AI-supported outcomes.

Beyond Personal Productivity — Redesigning Work with AI

AI can help people complete individual tasks faster, but the bigger opportunity for leaders is to rethink how work itself is designed. Instead of simply adding AI to existing workflows, leaders can decide which parts of a process should be automated, augmented, or kept human-led.

In this lecture, you’ll explore how AI can redesign recurring workflows, shift human effort toward higher-value work, and change the way teams operate across functions such as HR, finance, and customer service.

In This Lecture, You Will:

▸ Understand why leaders need to think beyond using AI for individual productivity and consider how teams and organizations should use AI.

▸ Recognize why simply adding AI tools to existing workflows may not deliver the expected productivity gains.

▸ Explore how redesigning a monthly business review can shift work from information processing toward interpretation, risk identification, recommendations, and organizational action.

▸ Identify repetitive steps that may be suitable for AI automation, including activities with consistent inputs, processes, and outputs.

▸ Recognize analysis-oriented steps where AI can accelerate thinking by synthesizing information, identifying patterns, and surfacing options while humans determine what the information means and what to do.

▸ Identify steps that require human judgment because context, relationships, and the consequences of errors are significant.

▸ Apply the question, “Which parts of this process belong to AI and which parts belong to humans?” when redesigning a recurring team process.

▸ Examine how AI can handle activities such as application screening, skills matching, interview scheduling, and routine candidate questions while recruiters focus on relationships, leadership potential, cultural fit, and final hiring judgment.

▸ Explore how AI can support finance through invoice processing, expense classification, standard reporting, and anomaly detection while finance leaders focus on risk management, resource allocation, strategic planning, and investment decisions.

▸ Understand how AI can handle routine customer service inquiries, request routing, and self-service support while human teams focus on complex, sensitive, and escalated situations.

▸ Recognize how the nature of human work can shift toward judgment-oriented activities rather than simply disappearing as AI takes on more routine work.

▸ Examine how AI agents can perform multi-step activities autonomously and why leaders still need to determine how much authority an agent should have.

▸ Apply the AI delegation framework to organizational workflows by considering what can be fully automated, what should be augmented with human involvement, and what should remain human-led.

▸ Redesign a recurring team process by identifying at least one activity for full automation, one for significant AI augmentation with a human in the loop, and one that should remain firmly human-led.

▸ Recognize why the greatest opportunity from AI is not simply doing existing work faster, but redesigning work so humans can spend more time on strategy, creativity, judgment, relationships, and consequential decisions.

Who This Lecture Is For:

• People Managers — redesigning team workflows so AI supports higher-value human work.

• Team Leaders — deciding which parts of recurring processes should be automated, augmented, or human-led.

• Business Leaders — thinking about AI adoption at the workflow and organizational level rather than only individual productivity.

• Functional Leaders — applying AI to recurring processes across areas such as HR, finance, and customer service.

Why This Lecture Is Useful for People Managers

People Managers influence how work is structured and how their teams spend their time. This lecture helps managers move beyond simply introducing AI tools and instead redesign workflows so routine work can be handled by AI while people focus more on judgment, relationships, strategy, creativity, and decisions that require human ownership.

Case Study: Klarna's AI Customer Service

Klarna’s AI-powered customer service shows both the potential and the limitations of AI at organizational scale. The system handled millions of customer conversations, improved response times, reduced costs, and automated many routine requests. But the company later recognized that customer service involves more than operational efficiency.

In this lecture, you’ll examine what happened at Klarna and explore how leaders can decide where AI should handle customer interactions and where human involvement remains essential.

In This Lecture, You Will:

▸ Examine how Klarna deployed an AI-powered customer service assistant to handle large volumes of customer interactions.

▸ Understand the operational results of Klarna’s AI customer service, including faster responses, reduced costs, and automated handling of routine customer inquiries.

▸ Recognize why strong operational performance does not necessarily mean an AI implementation is delivering the best overall customer experience.

▸ Explore how Klarna later increased emphasis on human support for complex customer situations after recognizing that the customer experience had become too heavily optimized around efficiency.

▸ Understand why the AI itself was not necessarily the problem, even though the company needed to rebalance the role of human interactions.

▸ Identify the importance of trust, empathy, relationship management, complex problem solving, and customer retention in customer service.

▸ Distinguish between asking whether AI can perform a task and asking what outcome the organization is actually trying to optimize.

▸ Examine how optimizing only for speed can produce a different result from optimizing for trust, loyalty, customer relationships, and long-term customer experience.

▸ Apply the four-level delegation framework to customer service interactions, from full automation to human-led work with AI support.

▸ Identify customer service activities suitable for full automation, such as password resets, order tracking, account updates, and basic FAQs.

▸ Recognize situations where AI can handle the process while humans review exceptions, as well as situations that should remain human-led with AI support.

▸ Identify escalated complaints, sensitive customer situations, high-value client relationships, and complex problem resolution as examples where human judgment remains important.

▸ Understand why successful AI transformation is not about replacing customer service or maximizing automation, but about deciding where automation creates value and where human involvement remains essential.

▸ Recognize how combining AI efficiency with human judgment can produce better outcomes than relying exclusively on either approach.

Who This Lecture Is For:

• People Managers — deciding how AI and human teams should work together in customer service.

• Customer Service Leaders — determining which customer interactions can be automated and which require human involvement.

• Business Leaders — evaluating AI transformation beyond efficiency and automation metrics.

• Team Leaders — applying delegation levels to customer-facing workflows.

Why This Lecture Is Useful for People Managers

People Managers need to balance efficiency with the quality of human interactions. This case study shows why the goal of AI implementation should not be maximum automation, but better outcomes—using AI where it creates value while preserving human judgment, empathy, and relationship management where they matter most.

The Keep / Change / Remove Framework

When AI changes how work gets done, leaders need a practical way to decide what should remain human, what needs to be redesigned, and what can be removed from manual work. The Keep-Change-Remove framework provides a simple way to make those decisions task by task.

In this lecture, you’ll learn how to apply the framework to accountability, judgment, relationships, processes, roles, and repetitive tasks, while also examining why some tasks that appear suitable for automation can contain hidden risks.

In This Lecture, You Will:

▸ Understand the Keep-Change-Remove framework and how it sorts tasks, roles, and processes into three categories.

▸ Identify work that belongs in Keep, including accountability, ambiguous judgment calls, relationship trust, and ethical decisions.

▸ Recognize why these human responsibilities remain important even as AI capabilities improve.

▸ Examine how Keep applies across functions such as HR, Marketing, Sales, Operations, and Finance.

▸ Identify work that belongs in Change, including approval structures, role scope, and reporting lines that may need to be redesigned as AI changes how work is performed.

▸ Understand how leaving existing processes unchanged while tasks and work volumes change can create bottlenecks.

▸ Identify work that may belong in Remove, including purely repetitive and rule-based tasks such as manual data entry, first-pass document sorting, routine scheduling, and basic data reconciliation.

▸ Recognize why removing a task from a person's responsibilities does not necessarily mean removing the person, as their time can shift toward judgment, coaching, and relationship work.

▸ Apply the three framework questions to determine whether a task should be protected, redesigned, or removed.

▸ Evaluate whether a process still makes sense after AI takes over part of the work and identify whether its approval structure, role definition, or reporting line needs redesigning.

▸ Recognize the risks of moving tasks into the Remove category too aggressively in pursuit of quick efficiency gains.

▸ Examine how seemingly routine tasks can contain hidden judgment, quality checks, relationship considerations, or fairness risks.

▸ Analyze the Amazon resume-screening example to understand how historical hiring data contributed to biased AI recommendations and why the tool was ultimately scrapped.

▸ Recognize why tasks involving hiring, performance, and other employment decisions require particular attention to hidden bias and fairness risks.

▸ Apply an additional risk question to people-related decisions: whether getting a task wrong could quietly disadvantage a group of people in a way that may not become obvious until later.

▸ Use the Keep-Change-Remove framework task-by-task rather than applying a blanket decision to an entire role.

Who This Lecture Is For:

• People Managers — deciding which team responsibilities should remain human-led, be redesigned, or be removed from manual work.

• Team Leaders — evaluating tasks and processes as AI changes how work is performed.

• HR and People Leaders — considering fairness and hidden risks when redesigning people-related work.

• Business Leaders — making practical decisions about organizational work redesign and AI adoption.

Why This Lecture Is Useful for People Managers

The Keep-Change-Remove framework gives People Managers a practical way to evaluate their team's work at the task level. It helps protect areas that require human accountability and judgment, identify processes that need redesign, and recognize repetitive work that can potentially be removed—while avoiding automation decisions that overlook hidden judgment, fairness, or relationship risks.

Building Readiness — Process, Timeline, Stakeholders (Team-Scoped)

Knowing what needs to change is only the beginning of an AI-driven redesign. Making that change work in a real organization requires a structured process, realistic timelines, and clear ownership across the people involved.

In this lecture, you’ll learn a six-stage process for moving from diagnosis to ongoing governance, understand realistic timelines for pilots and broader redesigns, and identify the stakeholders responsible for initiating, implementing, and owning the outcome.

In This Lecture, You Will:

▸ Understand the six stages of an AI-driven redesign: Diagnose, Pilot, Redesign Roles and Workflow, Reskill, Scale, and Govern.

▸ Apply the Diagnose stage to map where organizational design has drifted from how work is actually being performed.

▸ Understand why a Pilot should focus on one team, workflow, or specific process before expanding a redesign more broadly.

▸ Recognize how pilots can reveal resistance, AI performance gaps, and unexpected workflow bottlenecks while the cost of mistakes remains contained.

▸ Use evidence from a pilot to inform changes to job scopes, approval chains, and reporting lines during the Redesign Roles and Workflow stage.

▸ Recognize why Reskill is essential when redesigned roles require people to develop new capabilities, including critically directing and reviewing AI output or learning to coach rather than supervise.

▸ Understand how Scale expands a refined redesign from the pilot team to the wider function or organization.

▸ Recognize that Govern is an ongoing process for checking whether new decision rights, workflows, and roles are working as intended and adjusting them when necessary.

▸ Understand how the six-stage process can complement established change management approaches such as Kotter's 8-step model and ADKAR while remaining focused on decision rights, roles, and workflow.

▸ Distinguish between realistic timelines for a narrow pilot and a broader structural redesign.

▸ Recognize that a well-scoped pilot can often show results within 4–6 weeks, creating visible early wins and building confidence in the broader effort.

▸ Understand why a full structural redesign involving roles, reporting lines, and decision rights can realistically take 6–18 months, depending on organizational size and the number of functions involved.

▸ Identify the three distinct roles involved in a redesign: who initiates it, who implements it, and who is accountable for the outcome.

▸ Understand why senior leadership or a transformation office typically initiates the redesign as a strategic decision about organizational direction and priority.

▸ Recognize the shared implementation responsibilities of HR, function heads, and middle managers.

▸ Understand why assigning a single executive sponsor as the accountable owner helps maintain momentum when the redesign encounters friction.

▸ Identify the wider stakeholders involved, including leadership, HR, IT and data teams, legal and compliance, middle managers, and where relevant, employee representatives or unions.

▸ Recognize why involving relevant stakeholders at the appropriate stage is important for avoiding resistance and ensuring the redesign is properly supported.

▸ Understand why the success of a redesign depends not only on AI technology, but also on following the process in the right order, maintaining realistic timelines, and involving the right stakeholders at the right stages.

Who This Lecture Is For:

• People Managers — translating AI-driven redesign into how their teams work day to day.

• Team Leaders — participating in pilots, workflow redesign, reskilling, and scaling efforts.

• HR and People Leaders — supporting role definition, reskilling, and organizational redesign.

• Functional Leaders — applying redesign decisions to the realities of their specific functions.

• Business and Transformation Leaders — initiating, sponsoring, and coordinating broader AI-driven redesign efforts.

Why This Lecture Is Useful for People Managers

People Managers play a critical role in turning organizational redesign into practical changes in day-to-day work. This lecture gives managers a structured view of the redesign process, the time involved, and the stakeholders they need to work with so changes are tested, supported, and implemented rather than simply announced.

Cost-Benefit — Making the Business Case

Any AI-driven redesign eventually faces a practical question: What does it cost, and what do we get back? A credible business case needs to account for both sides honestly, including short-term costs, measurable benefits, and longer-term risks and gains.

In this lecture, you’ll examine the costs and benefits of AI-driven redesign, work through a claims-processing example, and learn which metrics can be tracked together to evaluate whether the redesign is actually delivering value.

In This Lecture, You Will:

▸ Identify the short-term costs of AI-driven redesign, including AI tool licensing, reskilling time, temporary productivity dips, and change-management effort.

▸ Understand why ignoring short-term costs can reduce the credibility of a business case when actual implementation does not match initial expectations.

▸ Recognize short-term benefits such as early efficiency gains from narrowly scoped pilots and morale improvements from removing repetitive, low-value work.

▸ Examine how a claims-processing redesign can create measurable efficiency gains by changing decision rights rather than simply introducing an AI tool.

▸ Calculate how reducing processing time for 600 low-complexity claims from approximately 45 minutes to 15 minutes can free roughly 300 hours of work per month.

▸ Understand how freed-up capacity can be redirected toward higher-complexity claims and coaching work rather than simply eliminating the time from the organization.

▸ Identify longer-term costs, including over-automation, skill atrophy, and cultural resistance.

▸ Recognize how excessive automation can remove tasks that contain judgment that was not obvious initially, making later correction more costly.

▸ Understand how relying on AI for a skill over time can weaken people's ability to perform that skill when human intervention is eventually needed.

▸ Recognize how rushed or poorly communicated redesigns can create distrust that continues beyond the initial rollout.

▸ Identify long-term benefits including structural agility, reduced cost to serve, and increased capacity for higher-value work.

▸ Understand how redesigned organizations can become more adaptable by developing experience with changing decision rights and roles.

▸ Recognize how people's time can shift from routine execution and supervision toward judgment, coaching, and relationship building.

▸ Track four key metrics for evaluating a redesign: cycle time, error or rework rate, cost to serve, and employee sentiment.

▸ Understand why these metrics should be tracked together rather than individually when evaluating redesign performance.

▸ Apply the principle behind Goodhart's law to understand why optimizing for a single metric, such as cycle time, can create unintended trade-offs with quality or employee sentiment.

▸ Build a more credible business case by tracking results consistently from the pilot stage rather than relying only on projections.

▸ Balance short-term costs and benefits with long-term risks and benefits when presenting an AI redesign business case.

Who This Lecture Is For:

• People Managers — evaluating whether an AI-driven redesign is creating measurable value for their teams.

• Team Leaders — tracking the operational and employee impact of redesigned workflows.

• Business Leaders — building and presenting a credible business case for AI-driven organizational change.

• HR and Functional Leaders — evaluating the people, productivity, and cost implications of redesign.

Why This Lecture Is Useful for People Managers

People Managers need to explain not only what an AI redesign could improve, but also what it will cost and what risks need to be managed. This lecture gives managers a practical way to quantify efficiency gains, recognize longer-term risks, and track cycle time, quality, cost, and employee sentiment together when evaluating the real value of change.