AI Change Management Playbook

Buying the tools is the easy part, and on its own it changes nothing. AI only pays off when people genuinely change how they work, and keep changing it as the tools keep moving. That takes real, ongoing effort, learning, experimenting, rebuilding workflows, and effort like that can't be mandated. People put it in when they believe in the why, so when it's clear and feels like theirs, they become the rollout's biggest force. Start here: get clear on what AI is for in your organisation, and the story you'll tell about it.

Some questions for your leadership team to work through:

People build intuition for AI only by using it: trying things, getting them wrong, working out what helps. Go broad before you go deep. Early on, the win is simply people using AI widely and getting a feel for what it can and can't do; the deeper automation comes later. So treat that time as part of the job, and make it safe to fail, since what doesn't work still teaches people what AI can and can't do.

Saying you encourage it isn't enough, though. A team that's already flat out won't experiment if it just means more hours on the same workload; they'll quietly skip it. You have to make real room for it, and there are two ways to do that, which work best together:

• Dedicated, protected time: a regular slot people are expected to use for learning and trying things.
• Team leaders protecting it day to day, treating experimentation as real work rather than something to squeeze in around the edges. They're the ones who decide what their people actually have time for.

Leaders back this once they see the payoff for themselves, so make that case to them directly.

You're asking a lot of people: to learn, change how they work, and keep adapting faster than anything has demanded before. Whether that feels worth it comes largely from the top. People can tell the difference between a leader who genuinely believes in this and one who has signed a budget and called it done, and they respond accordingly. So the sponsor's real job is to make the commitment visible, and keep it that way.

Strong sponsorship invests in the right tools, protects people's time, experiments alongside everyone else, and keeps the message alive in public. Weak sponsorship approves the cheapest option, leaves teams to find the time, goes silent after launch, and treats it as IT's problem.

Without clear guardrails, people are left with a hard choice: avoid the powerful new tools in case they break a rule they didn't know existed, or push ahead anyway and risk exposing company data. The first stalls the experimentation this rollout depends on, and the second creates the very risk you were trying to avoid.

An over-cautious policy is expensive too. The tools, used well, make people dramatically more effective, so lock them out with heavy-handed rules and they'll reach for shadow AI anyway, which is a bigger risk now than ever because the tools are so good. Aim for policy that keeps people safe and still lets them move fast.

Which tools you put in whose hands is one of the biggest decisions in the rollout. Unlike most technology, the right AI tool can make someone ten times more effective, not ten percent. And which tool is best keeps changing fast: one that helps a little this month might do the whole task on its own a few months later. Old approval cycles can't keep up, so by the time a tool clears a long review, it has changed, pricing included.

Tracking the right things helps you understand where you are on the adoption journey, what's working, and where to focus next. Set it up before you start: capture a baseline now, and get clear on what you're hoping to see, so the picture means something later.

Set all of this up now, even if lightly, so you have a clear starting line. Then treat it as a living process: revisit what good looks like, and shift where you focus as the tracking shows you more.

Every rollout needs one person who owns it. Your Directly Responsible Individual (DRI), or program owner, is the one person (or small core team) accountable for making AI adoption happen. They connect the leadership vision to what actually changes on the ground, and they drive the change itself: owning the plan, keeping it moving, shepherding people through it. Without someone holding that, even a well-funded program slowly drifts. It's a real role that takes real time, not a side-of-desk task for someone who's already stretched.

A great DRI fast-tracks a sensible tool request, connects a stuck team to the right champion, and spreads a win from one team to the rest. A struggling one sits on requests, funnels everything through themselves, and lets good wins stay buried in one team.

Champions are where adoption actually spreads. People trust a colleague who's getting real value from AI far more than any message from the top, and we've seen a strong champion network do more for adoption than any mandate or all-hands. Aim for at least one in every team.

The best way to find them is often just to ask. The people who put their hand up tend to be the ones who genuinely want to help, so be clear about what they're signing up for: describe what the role really involves, and you'll attract real champions rather than AI enthusiasts. Usage data can flag power users worth a tap on the shoulder too, but the open, well-framed ask is what surfaces the best ones.

Open the month with a short heads-up, not the full launch. Have an exec explain why the company is doing this and what the next month looks like, then introduce the champions (and any partner helping you) and the part they'll play. The ask of everyone else is small: complete the capability survey, get to know your champion, and feel free to look around the platform and its learning resources. Be clear the real start is still to come, so people know what's happening without feeling it's already begun.

The survey is how you find out where the organisation really is, rather than where you assume it is. Send it to everyone and ask about skill and confidence, which tools they actually use (approved or not), how they use them, and where they'd like help. Give it a couple of weeks so the answers are honest and considered. What comes back sets the accelerator's focus, shapes the workshops, and becomes the baseline you measure against later. You built this survey in Ready the environment; here it goes to work.

This month is about the role, not the tools. The AI workshops come later, in the accelerator, where champions take part alongside everyone else. Through the change-management workshop and some self-serve learning, get your champions to understand what the role involves, how the accelerator will run, and what a healthy team looks like, so they can start reading the signals. Give them a first change-management challenge to build the habit, like asking their team how they already use AI and bringing back what they hear, which also makes them the person their team starts coming to.

Set a light, regular cadence with the DRI: one sync early, once they've started talking to people, and one near the end, so the DRI hears what they're learning and can adjust. They're being readied here, not set loose, and nothing goes public until launch.

Put your champions to work pressure-testing the setup. As they experiment and hunt for real use cases, they'll hit the rough edges: a tool that won't connect, a missing connector, an unclear policy, a stalled access request. This is also when you do the real integration work, wiring the tools into your data, systems and workflows, since the gaps show up fastest when people actually use them. Far better to find all this now than on launch day. Feed what they surface to IT and the DRI, fix it as you go, and have everything solid before you scale.

Don't let the organisation arrive to an empty tool. Use the month to set the workspace up so people land somewhere ready, with a few things already useful on day one. Shaped by what your champions are finding, that can mean:

• Building a handful of genuinely useful, org-wide skills or agents: a brand tone-of-voice skill, say, or a slide builder preloaded with your internal and external deck guidelines.
• Pulling your shared resources into one easy-to-find place, the brand kit, key guidelines, common SOPs, so people have something solid to build their own agents and automations around.
• Turning on the right connectors and permissions, and org-wide search if your tools support it, based on what champions hit in testing, so the tools come preloaded with real use cases.
• Standing up the sharing spaces early in the month: the champions' own channel, and the space the whole organisation will use to share experiments once the accelerator begins.

Get your champions saving their best automations and use cases as they go, too, so there's real, homegrown work ready to show on launch day, in the kick-off and in the channels.

In the final week, bring the survey results and the champions' read of their teams to the execs and key stakeholders, so leadership goes into the accelerator with an honest picture of where the organisation actually stands. Use it to tailor the accelerator: where to focus, which workshops to run, and the order you bring teams in. Then confirm everyone knows their part for launch day, so the big push starts clean.

The launch sets the tone for the whole month, so make it an event. This is the one session everyone attends. An exec opens it and makes the case, you unveil the name you've given the month (an "AI Month" with its own identity), and you lay out what's ahead: what's expected, the line-up of sessions, the weekly challenges, and the hackathon to finish. Be clear about what people should be able to do by the end, and spotlight the champions so everyone knows who their go-to is. People should leave genuinely excited, and clear on how to take part.

Hands-on workshops are the heart of the month. People learn a genuinely new skill by doing, not watching, and the sessions are where a lot of the excitement comes from. From what we've seen, nothing builds momentum like a room full of people building together, so run them live and in person wherever you can.

Workshops are only part of the month. Between them, everyone needs something to do at their own level, so the energy doesn't dip and no one feels left out. A few things keep that going:

• A weekly challenge or project to complete and share, ideally built on their own work so anyone can take part whatever their starting point, with a simple progress track alongside it so people can see how far they've come.
• A weekly all-hands where people show what they've built and how they're using AI. Seeing a colleague's win does more than any slide.
• Shared channels and a standing weekly office hours, run with your change partner, where anyone can drop in and ask an expert.

These are the common places people gather, share and get unstuck. The aim is simple: wherever someone started, there's always a next thing for them, and somewhere to share it.

An accelerator is a living thing, not a fixed plan you run to the end. Steer it week by week.

Finish with a hackathon, the moment everything comes together and the month's energy peaks. Run it over two weeks so people have room to do something real.

• Week one is for ideas: a kick-off to spark them, time to form teams, and champions and office hours on hand as people shape what they'll build.
• Week two is for building, then a session to finish up and present. Let people volunteer to show their work, or pre-pick a few standouts, and give out prizes across a handful of categories (most useful, most creative, best first-timer, biggest team impact, that kind of thing).

Aim for everyone to take part. You end the month with a wave of real, working projects, the proof that the whole organisation has moved, and the start of a culture that carries on long after.

The quickest way to lose what you built is to let the program quietly wind down, so keep it staffed and visible. The DRI keeps owning it, the champions stay active in their teams, and the channels stay alive with real conversation. Keep the rhythm of health checks and catch-ups going; it's how you notice a network going quiet before it fades. And keep your executives in view, still using AI themselves and still talking about where it's taking the company, because the moment leadership goes quiet, people read it as over.

Even your most capable people will be out of date within months, that's just the pace of the tools, so learning can't be a one-off. Give people access to high-quality, up-to-date learning resources they can self-serve from any time, tailored to the tools and the work they actually do. The best of these are interactive: people learn by applying AI to their own tasks and walk away with a real use case, which is what makes the time genuinely pay off. Run workshops on demand when a real need shows up, have the DRI hold regular drop-in sessions where anyone can bring a question however basic, and re-run the capability survey now and then to see how far people have come and where the gaps are now. The message stays the same: there's always a next thing to learn, and always somewhere to learn it.

The excitement from the accelerator fades on its own unless you feed it. Keep celebrating wins out loud, and keep a clear window open onto what people across the company are building, a colleague's clever use is still the best advertisement there is. Keep a community going where people share and help each other, and run a hackathon every so often to re-spark the energy and surface the next wave of ideas. The aim is for experimentation to feel normal and shared, not something that happened once in a big month.

As fluency spreads, the work shifts from using AI to building with it. Once the basics are second nature, the bigger gains move to more advanced ground: custom agents, automations running in the background, whole workflows rebuilt around AI. Back your most capable people, champions especially, to go there, give them the room and the tools, and make sure IT and policy keep pace as the builds get more powerful and start touching real data and systems. This is where AI shifts from helping people work faster to changing how the work gets done.

The tools keep racing ahead, so your setup can't stand still. Revisit your tooling, budgets, spend and access on a regular cadence, and keep asking whether you're still on the best option for each job, the answer changes often. As capability climbs, so should your sense of what "good" looks like; the bar you set six months ago is probably too low now. And keep checking it's actually paying off: move your measurement on from "are people using it" toward the real business impact it's creating, so you can see the return and keep investing with confidence.

Eventually this stops being a program and becomes simply how you work, and you can make that real. Build AI fluency into job descriptions and into how performance is talked about, so it's a genuine expectation rather than a nice-to-have. Hire for it: look for people who are already AI-native, and get clear as a leadership team on what that means for each role. And fold it into onboarding, so everyone who joins from here starts fluent. Once using AI well is simply part of the job, it's no longer a change you're managing, it's how the company works.