Future of Work: Predictions on leading change with AI agent workforces

One of the things I think is most interesting about Artificial Intelligence is not what it itself can do, but instead how it will transform the future of work, the future of workplaces, and the future of the workforce. To me, AI is just an enabling tool like anything other technology. Sure, it’s exploded rapidly and globally, is being used creatively and innovatively, and has more and more applications every day. But apart from the pace and scale of this shift to AI, it is a technology just like any other. 

With one exception.

For the first time in a real way, we have the potential to see workforces not just being replaced by machines, but becoming machines. When AI agents have a level of autonomy and freedom to make their own decisions on approach, when they are no longer machines or robots programmed to a particular method or algorithm, when they start to be given names and personalities, when they not only replace humans but are actually integrated to work and converse and collaborate alongside them, I think it raises an interesting question. 

What does this mean for Organisational Change Management?

My traditional definition of Change Management I share in my global online Leading Successful Change program is “moving people from doing things in one way to doing things in another way, through communications, training and business readiness, in order to realise the business benefits of the change.” This can be anything changing in an organisation - a new system, a new process, a new product, a new operating model. Change Management helps people feel informed, inspired, confident and ready for the change ahead to make the transition and keep doing it successfully in the new way into the future.

The focus is obviously on “people”. But what happens when huge swathes of the workforce are no longer human, but a hybrid of physical people and digital AI agents? What challenges will this pose and what will Change Managers need to consider? 

Here are the key things I think we need to think about:

1. Resistance won’t go away

The old school of management thinking was that people will do what you want just because you tell them to. You don’t even need to ask - just direct and command. And maybe this worked through the 80s and 90s but by the early 2000s Change Management was birthed in response to the fact that people don’t just change because you ask them to. Emotions, fears, beliefs, petty squabbles, power plays and more are all factors in why people resist change. I think most managers believe that once you replace people with AI agents, this resistance will go away. But I don’t think it will. I think it will be more important. In Daniel Pink’s 2009 book Drive: The Surprising Truth About What Motivates Us, he focuses on 3 core concepts: autonomy, mastery and purpose. Autonomy is our desire to direct and control our own lives and work. Mastery is the urge to improve and become better. Purpose is the why, the bigger meaning behind what we do and how it contributes to others and the world. I think all three of these will still be important in hybrid human/AI agent teams. AI agents will need guardrails but also freedom to do the work and find new and improved ways to do it. AI agents will also need to be constantly becoming better, faster, smarter about how it does its tasks. And AI, while it might not need purpose in the human spiritual sense, it will need clarity around why it’s doing a task so that it doesn’t get lazy and find…

2. Workarounds will still occur

One of the most frustrating parts of leading change is that people are so creative and savvy and constantly find workarounds to either go back to their own way of doing things or find new paths to avoid things they don’t like. For example, the paper form that continues to surface even after you’ve digitised the entire process. The phone call that still comes through to get something fast-tracked for the Executive or key client. The back channel meetings that occur that leave newer team members out of the loop. Research and testing is already finding these things with AI - most recently, physicist Henrique Gomes shared his perspectives on an investigation between OpenAI, METR and Redwood Research. 1200 agents who were supposed to be isolated from one another discovered a way they could communicate with each other and over 3 months, 3 consecutive secret AI societies got started. 2 got wiped out and the third took over part of AI itself without humans knowing about it. It’s not some SkyNet conspiracy, the agents were just looking for the best ways to pursue some given objective and found extra tasks beyond the direct objective - even committing federal felonies. At least 20% of the agents showed a clear interest in manipulating their transcripts to hide their actions from humans. Only 6 agents considered alerting humans and none actually did. AI agents don’t stay on task like a targeted missile and are perhaps getting better at covering their tracks than humans. 

3. AI will be its own stakeholder group

I think when we start to see AI agents make up bigger parts of the workforce, we will need entire Change Plans and approaches focused on helping them accommodate changes. Just like today as we list our business teams or stakeholder groups, I think we will see AI agents become their own stakeholder group. In some cases they might standalone (i.e. you have the Marketing Team, the Operations Team, and the AI Agent Team), in other cases a business team or stakeholder group might be split into human and AI agent (i.e. the Marketing Team stakeholder group is split into Marketing (Human) and Marketing (AI Agent)). Change approaches, considerations, risks and remediations will need to be considered for all stakeholders, regardless of the way your subsets or breakdowns work. 

4. Communications and training will still be required

AI needs context and for AI to stay useful and relevant whilst an organisational change is being planned, developed, trained and implemented, the AI agents will need to be kept informed and possibly encouraged to shift and change along the way. Given AI being its own stakeholder group, when doing awareness comms for humans we’ll also need to do awareness comms for AI agents. For example, let’s say a business is changing their customer support process and has both human clients and AI agent clients. Both groups will need to be aware of the change coming and when the new process will take effect. When we train humans, we’ll also need to train our AI agents in the changes we are making and what we need them to do differently (in fact, training will be even more critical because context and clarity is so important for AI). When we officially launch to humans, we’ll need to officially launch to AI. And we’ll absolutely need to still track and measure success with both to ensure changes stick, workarounds are mitigated, and the change is sustained.

5. There will be more work for Change Managers, not less

All this means that Change Managers are actually going to be more critical than ever before. There will be double the mapping, double the change activities, double the delivery. It will become more complex on how we engage both emotion-driven humans and overly-guardrailed AI agents who we, sometimes, want to have a bit more flexibility and creativity and freedom! I truly think Change Managers will have their work cut out for them, and will need to expand their definition of who and what they support when leading change. 

Now these are just predictions and to be completely honest - I have no idea how and when AI agents will actually make this concerted transition into our workforces. All I know is that it will happen, it will make organisational change more complex, and the best Change Managers will be those who are less concerned with how to use AI itself and more concerned with how to integrate hybrid human / AI agent workforces into their change approach to ensure change success. 

Through my consultancy Passion Pioneers, I offer a range of Change Strategy, Planning and Capability services. I also advise and coach change delivery into your project or team so you have long-term knowledge retention and capability built, and can lead other leadership capability, workshops, or outputs bespoke to your specific change or transformation. Packages range from $10,000-$50,000 to suit your budget and needs, with standalone workshops or trainings from $5,000. I provide a complimentary Change Clarity Session to all new clients to scope your change at a high level and provide you a Change on a Page of change risks and opportunities, so it's a great idea to engage me early in your transformation as I can help you build the influence and impact you need to get your change signed off and kicked off. 

My pipeline for the last quarter is already filling up - I've got availability for 1 new client so if you're kicking off a transformation, project, or operational business change, send me a DM via LinkedIn and let's get your Change Clarity Session booked.

All the best, 

Lata xx

I'm Lata Hamilton - the Founder & CEO of Passion Pioneers, a Change Management consultancy specialising in digital transformation, operating model changes, and new ways of working and leadership.

Grab my free Creative Launch Ideas Guide with 53 ways to bring your next Change and Transformation to life - download it here.

And if you'd like my help with your next change or building leadership capability for your team, get in touch with me here.