Agentic AI for HR Leaders (Inc. Copilot & Power Automate )

In this lecture, we discussed the foundations of AI Literacy for HR leaders at a high level. We started off by examining popular workplace narratives around AI replacing jobs and introduced a reflective question that we will revisit throughout the course. Then we explored essential terminology like AI, machine learning, deep learning, generative AI, agentic AI, and data science, followed by HR-specific examples showing how each type is used in functions like attrition prediction, resume screening, sentiment detection, and training content creation. Next, we examined a real-world case study on IBM’s AskHR chatbot, which combined generative and agentic AI to transform HR operations while also raising questions around media narratives and job displacement. Finally, we looked at the Gartner Hype Cycle to understand how technologies evolve from inflated expectations to maturity, helping leaders differentiate between hype and strategic value.

In this lecture, we discussed practical AI use cases in HR at a high level. We started off by exploring the most common HR challenges where AI is already making an impact, and then mapped real use cases across the employee lifecycle—from recruitment and onboarding to engagement, learning, and performance management. We discussed how generative and agentic AI can help with writing job descriptions, screening resumes, creating training material, managing internal communications, predicting attrition, and more. Then we looked at how these use cases actually work behind the scenes using techniques like large language models, self-supervised learning, prompt engineering, and retrieval-augmented generation. Finally, we looked at the broader ecosystem of agentic AI providers and walked through real-world case studies from companies like Chipotle, Ubisoft, and Cognizant to show how agentic AI is already being used in live enterprise settings to speed up hiring, improve onboarding, enhance feedback, and automate policy queries—reinforcing that AI is not here to replace HR, but to help us focus on more strategic and human-centric work.

In this lecture, we discussed how to practically bring generative and agentic AI into your existing HR systems at a high level. We started off with a quick recap of what we have covered so far, including key terms, tools like the AI Hype Cycle, use cases across the talent lifecycle, and the four major categories of agentic AI providers. Then we discussed the three most common categories of HR use cases—content creation, research and analysis, and routine workflow automation—along with hands-on examples such as job description drafting, onboarding deck creation, salary benchmarking, and attrition trend analysis using tools like Microsoft 365 Copilot. Finally, we looked at how legacy HR platforms like Workday, SAP, Oracle, and Zoho can already be integrated with Microsoft 365 Copilot using pre-built connectors and digital agents, enabling seamless access to enterprise data and automation within the Microsoft ecosystem.

In this demo, we looked at how Microsoft 365 Copilot can streamline the hiring process at a high level. We started off with the challenges faced by HR teams in writing inclusive job descriptions, manually reviewing resumes, and ensuring alignment with organizational values. Then we discussed how Copilot in Word helps refine a rough job description using prompts, and how the Analyzer Agent in Copilot Chat can compare candidate CVs against the job description using structured tables and output recommendations. Finally, we looked at how Copilot in Outlook can draft a personalized offer letter, completing the hiring workflow with professional communication. Throughout the session, we also applied prompting best practices such as setting the persona, giving context, uploading reference documents, and defining structured outputs to ensure Copilot generated meaningful and actionable results for HR tasks.

In this lecture we looked at how to create a custom agent in Microsoft 365 Copilot Chat at a high level. We started off with an overview of different methods available to build agents including using ready-made templates, describing the agent in plain English, and configuring one from scratch. Then we discussed how to apply prompt engineering best practices by clearly defining the agent’s purpose, tone, response format, and fallback rules. Finally, we looked at how to build a leave policy Q&A agent using an internal SharePoint document as the knowledge source, walked through the manual configuration steps including adding rules and sample prompts, and tested the agent with both valid and invalid queries to demonstrate accurate response handling and safe refusal.

In this lecture we looked at how to automate a leave approval workflow using Microsoft Power Automate and Forms at a high level. We started off with a common HR challenge where leave requests are manually tracked through emails or messages, often leading to delays and errors. Then we discussed how Power Automate can streamline this process by connecting various Microsoft 365 apps and eliminating the need for manual follow-ups. Finally, we looked at how to build the actual workflow step by step—from creating a form trigger, extracting response details, setting up an approval process, and sending dynamic notifications based on the manager’s decision—all of which makes the leave approval process faster, more consistent, and employee-friendly.

In this lecture we looked at Responsible AI in HR at a high level. We started off by recapping our journey so far and discussed how AI is becoming more powerful and autonomous, which also increases its risks. Then we explored the differences between key terms like AI governance, responsible AI, ethical AI and trustworthy AI, and broke down the four pillars of AI ethics—risk, harm, bias and fairness—using relatable HR examples. We also discussed various types of risks such as discrimination, privacy violations, misinformation and misuse, and looked at different types of harm to individuals, organizations and ecosystems. After that, we understood the difference between model bias and ethical bias and studied types of fairness including group, subgroup and individual fairness. Finally we looked at the NIST characteristics of trustworthy AI through the CV screening use case, followed by a brief overview of global AI regulations such as the GDPR, EU AI Act and US state laws, so that you are now equipped with both ethical and legal lenses for using AI responsibly in HR.

In this lecture, we discussed how to design an AI literacy program for your team at a high level. We started off by revisiting what we’ve covered so far—from AI basics to advanced use cases and responsible adoption—even within legacy HR systems. Then we explored why AI literacy is critical today, especially in the face of job losses driven by automation and the rising misuse of AI tools without proper guidance. We looked at what AI literacy really means using the EU AI Act definition, and broke it down into skills, knowledge, and understanding with relatable examples. Finally, we looked at a step-by-step approach to designing an effective AI literacy program within your organization, starting from assessing current levels to developing role-based modules, delivering training, and ensuring continuous improvement.