AI in Change Management: Generative vs. Agentic AI and What’s Next
As organizations continue moving through rapid transformation, change managers are expected to deliver faster, smarter, and with more precision than ever before. That’s why it’s critical for both practitioners and clients to understand the AI tools already available, and the ones emerging just over the horizon. From Generative AI to Agentic AI, we’re entering a new era where technology can accelerate project delivery, free up capacity for strategic work, and reshape how we partner with teams, explains Moshe Cole, senior strategist at one of the nation’s largest and reliable public utilities. But with that opportunity comes new questions about trust, data, integration, and the irreplaceable role of humans in the loop.
Generative AI in Change Management: Today’s Most Accessible Accelerator
Most change practitioners are already familiar with Generative AI, even if only at a basic level (think ChatGPT). At its core, Generative AI is a content engine. Give it the right inputs such as project materials, sanitized data, and structured documents, and it can produce drafts, summaries, emails, FAQs, and training content in seconds. But to use Generative AI effectively, it’s important to understand how the content is being generated. Models draw from both what you upload and what they’ve been trained on. That means practitioners need to watch for inaccuracies or “hallucinations” and know how to validate output quickly. When used carefully and intentionally, it becomes a powerful assistant rather than a rogue content generator that you must babysit. One of the easiest analogies is something that most consultants know well: travel. You can tell a generative model, “I’m going to Minneapolis on these dates—give me an itinerary and show me flight options,” and it will do that instantly. But there’s a limit: it won’t book your trip. That’s where the next evolution of AI comes in: Agentic AI.
Agentic AI and the Future of Change Management
Agentic AI takes everything Generative AI does and adds action. Instead of stopping at suggestions, Agentic AI can make decisions and execute tasks within parameters that you set. While Generative AI can recommend flight options, Agentic AI can book the flights, reserve your hotel, and arrange your rental car. You simply give it criteria such as dates, price limits, and seating preferences, and then it acts.
From a change management perspective, this unlocks enormous potential. Most change managers have acted as a project’s “inbox agent,” someone who reads incoming project questions and drafts appropriate responses. Agentic AI can take care of the inbox agent’s role. That kind of automation could reduce inbox load and free change practitioners for more strategic, human-centered work. In fact, Cole recently built an inbox agent for a client that successfully answered about 25% of standardized project questions. The model performed well because it was trained only on clean, controlled, and approved content so responses were accurate and consistent. However, because the client wouldn’t allow integration with Outlook due to privacy concerns, the AI couldn’t act autonomously. It had to be manually copied and pasted. This is a perfect example of where we are today: the technology is ready, but adoption and trust are still catching up.
The Lag: Trust, Data, and Integration
Clients, especially those in regulated industries like aerospace, defense, government, and finance, naturally hesitate to give AI tools access to data, systems, or login credentials. And they’re right to pause.
Questions emerge quickly:
How is the AI accessing my accounts?
Where is my data stored?
Can the tool act without my knowledge?
How do we ensure privacy, compliance, and accuracy?
These concerns are not small hurdles. Without system integration, Agentic AI can’t function at full capacity. And without clean, high-quality data, even the smartest AI underperforms. It’s no surprise then that 95% of AI projects have not reached ROI expectations. Not because the technology is flawed, but because organizations are still learning how to prepare, pilot, and implement AI. For companies just starting their AI journey: you’re not behind. Everyone, from small teams to global enterprises, is still figuring this out.
Where to Start: Practical Guidance for Teams and Leaders
If you’re exploring AI for change work, consider these steps:
Start with a clear use case.
Don’t pursue AI because it’s trendy. Identify a repetitive, content-heavy, or rules-based workflow that bogs down your team.Review the end-to-end business process.
Understand how work actually gets done. AI will highlight inefficiencies—but it won’t fix a broken process.Evaluate tools already on the market.
There are affordable, streamlined solutions available now that don’t require building a custom Large Language Model (LLM).Prepare your data.
AI is only as strong as the data it’s fed. Clean, structured, accurate data is non-negotiable.Keep a human in the loop.
AI can move fast, but it cannot do empathy. Adoption challenges, resistance, emotions, and cultural dynamics still require human change practitioners who can listen, contextualize, and guide.
The Bottom Line: AI Accelerates but Humans Enable Adoption
AI can speed content creation, automate tasks, and optimize workflows. But it cannot build trust. It cannot sense hesitation in a stakeholder’s voice. It cannot influence, reassure, or inspire. That’s where change managers remain essential. The future of change work isn’t human vs. AI, it’s humans who know how to use AI well. And as the technology continues to mature, the organizations that benefit most will be those that pair smart tools with skilled, empathetic practitioners who understand both processes and people.
Contact ChangeStaffing for support with AI change readiness and adoption within your organization.
Thank you to Moshe Cole for his thought leadership and for collaborating with us on this blog.
Written by Kylette Harrison.