A Point of View on AI, Change, and Change Management

As Prosci’s Chief Innovation Officer, I do not claim to be an expert in AI nor a predictor of the future. But I am paying attention to the conversation right now. And I’ve spent two decades studying change – how it happens, why it’s a challenge, and what variables contribute most to successful change. This article synthesizes some of my perspectives as an active, engaged listener who has also been a keen observer of organizational behavior, change approaches, and keenly watched a recent radical change response and reaction in the RTW / Future of Work space. Here are a few perspectives about generative AI’s impact on change (generally), change management (as a discipline), and change practitioners (as users of the capabilities unfolding every day).

Introduction

I still remember the Loom video – a narrated screen capture of a browser pointed at a url I had never heard of (https://chat.openai.com/chat) by Paul Gonzalez, Prosci’s digital product lead, – and his opening line, “I don’t know that I have ever been more excited about the future and more fearful of my own relevance.” Paul then proceeded to demonstrate the power of the recently released chatGPT 3.5 with some (what I now realize were) simple prompts around Prosci’s own change model within a given context that proceeded to yield jaw dropping results, word… by… word. He added a few more well-crafted and ingeniously intriguing examples within our domain and body of work referencing our models and content. I sat in stunned wonder.

Now, I imagine many of your first exposure to Chat GPT was to something sillier and more playful. “In the voice of Hunter S. Thompson, provide a play-by-play of a 2-hour heated PTA meeting,” was a classic one of my colleagues shared. “Write me a NYC style rap about grocery shopping for Thanksgiving.” Or how about starting to mesh the creative with the relevant using, “Write me a six-stanza poem about resistance to change in a hospital deploying electronic health records.” Even with the silly prompts, we watched how high probability statistics guessing the next word (or pixel) could yield incredibly humanistic and realistic results.

I began using Chat GPT to conduct some of my own analysis. In a recent webinar, several hundred attendees anonymously answered the question: what services does your CMO provide to the change practitioners and agents in the organization? I had already conducted my own affinity analysis and synthesis, pattern excavation, and summation of this particular stack of data. But, I wondered how this new sense-making tool might address it. I put the stack in and took a few different approaches with the prompt. Summarize in the tone of HBR. Identify the top ten themes with descriptions and examples. Summarize the top seen themes as action items. I “knew” what story the data had to tell having already explored it, so to see the AI-generated outputs was both eerily familiar and a brand-new experience, exposing broad expanses of possibility. I used this strategy to extract patterns, summations, and starting points from a variety of answers to thoughtful questions I had anonymously asked at scale.

A second opportunity for personal productivity application came a few weeks later. I had commissioned a side research project with the support of one of our amazingly talented Training Success Specialists. To satisfy a curiosity about the emergence of the “Chief Transformation Officer” role, we devised an approach to build a data set of organizations and newly appointed CTOs. Our data set aimed at exploring where these CTOs came from – internal vs external, previous role, background, etc. – and what they were being charged to accomplish – reporting relationship, staff or functional oversight, mandate/charter, etc. We had a nice data set and had started to generate some interesting analysis. But at one point, I started to wonder how different dimensions might enrich the analysis. Might patterns emerge if I added a column for industry and looked at the data through that lens (digression: banking, financial services, and health care came in at the top). Do CTOs show up in smaller, mid-sized, or larger organizations? Does the age of the organization impact the creation of a CTO role? I knew I could find and transcribe the values into the cells of my table to complete it – the data was publicly available, and I could easily gather it all… with time. Instead, I leveraged a few prompts and generative AI to complete the data set in a matter of minutes, literally.?

I have been listening to the podcasts I see in my the feed that have a bent toward AI generated opportunities . HBR IdeaCast has presented a balanced view on the potential issues and impacts of generative AI on productivity, creativity, culture, and strategy. Josh Bersin has explored the impact on human capital, talent, and HR technology. Scott Galloway has explored all sorts of dimensions of digitization and what it means to jobs and people. I’ve been looking for the patterns and seeing what lenses will help me make sense of what’s happening right now and what is in store. I can guarantee it’s not right, but it’s my read right now.

AI and Change

Before exploring how AI will impact both the discipline and the practitioners of change management, I’m going to start with a few higher-level observations about AI and change. For over 25 years we have been studying how to support people through change, and you cannot do that without learning a bit about organizations, changes, and the dynamics of change. With that as a backdrop, here are the four major disruption points I’m paying attention to on the “AI and disruptive change” landscape.

Job Disruption

Job loss is certainly one of the two biggest fears I’ve often heard expressed related to AI. However, while listening to an HBR IdeaCast I heard a quote that captures my view on job disruption along the lines of “Robots will not replace people. But the people using robots will displace the people who aren’t using robots.” Certainly, jobs will change, how could they not? Consider for a moment: what tasks do I complete on a weekly basis that could be made better, faster, more effectively with the support of generative AI? The automation of tasks can result in some jobs not being needed. Go ahead and ask any employees in the agricultural or manufacturing sectors who experienced waves of automation decades earlier. According to a World Economic Forum report the four fastest declining jobs were expected to be Bank Tellers and Related Clerks, Postal Service Clerks, Cashiers and Ticket Clerks, and Data Entry Clerks. These jobs are impacted by the digital transformation and the additional possibilities available once most of the objects of their work were digitized. On the flip side, a number of jobs are expected to grow over the next four years including AI and Machine Learning Specialists at the top of the list, followed by Sustainability Specialists, Business Intelligence Analysts, and Information Security Analysts. These jobs center on bringing forward the value of a digitized and connected world.

Interestingly, the concern around job disruption and loss is occurring against a backdrop of leading organizations moving toward an orientation around capability and skill, rather than simply the job and its title. This orientation to the people that make up the organization may prove a good springboard for incorporating generative AI-assisted capabilities and skills.

Industry Disruption

There will be a spectrum of industry disruption based on how much of the tasks and work of an industry overlaps with the features and capabilities of AI. A recent Accenture report referenced by the World Economic Forum stack-ranked industries based on the anticipated disruption given the proportion of “non-language tasks” as compared to tasks that could be or were likely to be automated by AI. Banking, Insurance, Software, and Capital Markets top the list as most disrupt-able, while Natural Resources, Chemicals, and Customer Goods & Services land at the end of least disrupt-able.

This is considering the collection of tasks that make up the values and services provided by a given industry. Even for organizations from the least disrupt-able industries, there will likely be significant disruption within their walls with their language-heavy tasks, and also from the vendors they use.

Organizational Disruption

One step more granular than industry disruption would be organizational disruption. Within the organization, change will come with various motivations. A number of AI-driven changes will be aimed at efficiency – helping the organization and employees do what they already do, better. Other AI-driven changes will be aimed at growth – helping the organization to expand and explore new opportunities. And still other AI-driven changes that organizations experience will be transformational – helping the organization reimagine itself, its operations, its services, and its value.

All three of these types of changes – efficiency, growth, transformation – will likely be occurring concurrently and be intertwined. The nature of the technology and its reach lend itself to interconnectedness. But, leaders of organizations need to establish focus and priority around which types of change will be addressed by who and when.

Organizations are also likely to experience second wave AI disruptions inside their own walls, brought through the door by vendors, suppliers, or customers. Around 2011, when our founder was reluctant to step into social media, a wise mentor once told me, “Prosci and ADKAR are on Twitter – it all depends on if we want to be part of the conversation.” I feel a parallel here. You may be a small business with few augmentable tasks in your business operations, but you are using a popular online accounting solution to keep your books, and you may be surprised to find a bot or copilot chat window as the only thing you see on the screen when you log in next month. AI will have an impact inside your walls, whether you invite it in or not. The better path is to be thoughtful and ready with how to provide your people with the guardrails and rumble strips they need to effectively and safely experiment with how to improve their own work process and output.

Task Disruption

Task disruption is likely where much of the efficiency gains will be coming. Encourage employees to think about how they can “do more or do better” by flexing the capabilities of AI in a responsible and safe manner. What do I do daily or weekly that AI can help me with if used efficiently – better, faster, more effectively?

There are several learning curves here. First, learning which tasks overlap with what AI is good at. Next, learning how to best interact with the interface to get the best outputs back. “Garbage in, garbage out” is not going away regardless of how powerful the algorithm is (garbage prompts yield garbage outputs). Finally, learning what is okay to do with the outputs, and what is not okay to do. Sounds simple, but this is a learning journey that all employees, top to bottom, in all organizations, left to right, will be taking in coming years.

AI and Change Management Discipline

There are several ways I see AI impacting the discipline of change management. At this level, we don’t yet know how AI impacts the “day job” of the change practitioner, but we are looking at the overall supply and demand for the services and value provided by the discipline of change management. To ensure we are at least on common ground, if not agreement, on that point, I would put forward “preparing, equipping, and supporting people through the changes they experience at work so that they themselves, the projects, and the organization are more successful” as the purpose of change management and what we’ll anchor to in terms of the collision of AI and the CM discipline.

Change Management Demand Generation

The changes listed in the “AI and Change” section – job disruption, industry disruption, organizational disruption, task disruption – all will trigger numerous changes in organizations that will need to be better managed (i.e., need some degree of change management applied). Whether the organization is using AI to power a new product for clients, to improve sales enablement, to generate social media content, or to gain efficiencies, people will need to adopt new behaviors and demonstrate new capabilies as their daily tasks are impacted. Projects, initiatives, transformations will be wave upon wave of change dependent on employee adoption. Demand for “preparing, equipping, and supporting people through change” will increase as organizations explore the changes needed to incorporate AI into operations and experimentation.??

Change Management Purpose Displacement

The ultimate value created by change management comes from “preparing, equipping, and supporting people to adopt a change to the way they do work.” If AI removes some of that “change journey” – similarly to how intuitive interfaces reduce some of the “journey” of navigating an application – the AI might eliminate some of the root “people side change” required by projects and initiatives. This is a much longer term view, several generations of disruption out, but it is on the horizon of possibilities to examine.

AI and the Change Management Practitioner

AI and the CM Practitioner is the final layer of impact to explore. A few initial thoughts on how AI will impact the work of the change practitioner.

First, there has been some great conversation already in the industry around how change management practitioners might use AI in their own work. Dan Olson from ChangeGuild facilitated what I hear was a great session in the Twin Cities for MnCMN on Using ChatGPT in your Change Practice. ?Dr. Kate Bryne with EverChange in Australia delivered an engaging podcast and a primer deck providing a great overview, guidelines, prompt formulas and guidelines, and accessible use cases.

This impact is tied to the “task disruption” for the CM practitioner themselves. “Task disruption analysis” begins with a clear frame for understanding the tasks of the change management practitioner. This is different than the methodology or toolset followed, it’s a description of the action steps taken to advanced change work on an initiative. A simple starting point might be: engaging, assessing, planning, activating, measuring. The question becomes, for example, what elements of my daily work engaging sponsors and the project manager might generative AI support me with? Maybe it is a first draft of a letter, or an initial agenda of a kickoff meeting in table format complete with activities. To build a a foundation, Prosci has an open research study exploring "The AI Effect and Change Management" that you can participate in right now.

There is still tremendous ground to be explored related to which change practitioner tasks most overlay with the capabilities of AI, as described above. But, to conclude this initial reflection, I’d leave you with one of my favorite takes on generative AI. Tim Ferriss interviewed Rick Rubin and asked about the impact of AI on creative work. Rubin’s response was along the lines of “As an ends; not interesting. As a means; very interesting.” He proceeded to use the analogy of a hip hop producer crate diving. Digging through tape after tape after tape, the producer is not looking for the next hit song or breakthrough artist. The producer is looking for moments that inspire and enable the creative to connect and see their work in a different light. As both a change practitioner and a user, I’ve found this “as a means; very interesting” to be a good description of the application as a source for crate diving, increasing the sample size of potential moments.

Conclusion

If someone told me they know how AI will unfold, I’d probably stop listening. The pace of development is lightning fast. From what I can tell, the application horizon continues to grow even as we work to wrap our hands and heads around it. From where I sit, I think there is massive opportunity for:

• Better starting points
• Quicker sense making
• Increasing sample sizes

In various tasks, jobs, and industries – and when applied to efficiency, growth, and transformation efforts – these capabilities can advance progress and unlock potential. I'll be watching, just like you, to see how these waves of disruption and opportunity will wash over teams and organizations.

Final thought. Conversation is really the most fundamental “user interface” there is for conveying information – beginning with story telling. But interacting/conversing with digital data has taken many forms. We began interacting with that data first through punch cards, then through keyboards, then GUIs, and then our access to data extended beyond our own computer and to the entire network of data through access to properly structured data sets if we know the correct and specific syntax for asking. With generative AI, we have reduced the barrier of access to information to the most basic and simple format – conversation. It poses a fascinating next chapter in the evolution of accessibility to information and all the woes and benefits that access may bring.

#####

Tim Creasey, Chief Innovation Officer at?Prosci, is a researcher and thought leader on managing the people side of change. His work forms the foundation of the largest body of knowledge in the world on change management. He has spoken to hundreds of thousands of people across the globe, helping them to see how successful change is unlockable when we prepare, equip, and support people through change.