The short answer

Change management should start at least six months before go-live — ideally at the point the technology decision is made. The behaviours, trust, and awareness required for successful adoption cannot be built in weeks. They require sustained effort across three distinct phases, each with a different purpose.

The problem with late change management

Late change management is not just suboptimal. It is structurally incapable of achieving what adoption requires.

Consider what successful adoption actually demands. People need to understand why the change is happening — not just what is changing. They need time to form a view on it, ask questions, and arrive at a position of acceptance before the pressure of go-live arrives. Frontline managers need to understand the change well enough to handle the questions their teams will bring them. Resistors need space to surface concerns and feel heard.

None of these things can happen in a four-week pre-launch window. Not because people aren't willing, but because they take time. Trust takes time. Familiarity takes time. The shift from "this is happening to me" to "this is something I'm choosing to engage with" takes time.

70%
of transformations that fail cite people and change issues as the primary cause
more likely to meet objectives when change management is excellent vs poor
6mo
minimum lead time for meaningful change management before a major go-live

When change management starts late, you see predictable consequences. Resistance that could have been addressed in structured conversations becomes entrenched opposition by go-live. Training that people can barely absorb under normal conditions becomes impossible to absorb when they're also managing the anxiety of a live system. Managers who haven't been prepared become unhelpful at exactly the moment their teams need most support.

The go-live becomes a crisis. And in the aftermath, adoption rates stall, workarounds proliferate, and within six months a significant portion of the organisation has quietly reverted to old ways of working.

The six-month change management timeline

What does it actually look like to start six months out? Here is the framework we use with clients across AI and digital transformation programmes.

Months 6–4
Foundation

Build awareness and understand your audience

This phase is about understanding the landscape before trying to change it. Who will be affected, and how? Where is the resistance likely to come from, and why? What are the competing priorities that will make adoption harder?

  • Stakeholder mapping — identify every group affected and their likely response
  • Impact assessment — document how roles, processes and day-to-day work will change
  • Manager engagement — brief all people managers before any broader communication
  • Early communications — explain the why, not just the what. Be honest about implications
  • Identify change champions — find early adopters who can become peer advocates
Months 3–2
Preparation

Equip people and address concerns

This phase moves from awareness to capability. People know the change is coming — now they need the knowledge, skills, and confidence to engage with it. This is also when concerns surface most visibly, and when listening matters as much as communicating.

  • Role-specific training begins — not generic, but tailored to each affected team
  • Hands-on access to the technology — people should touch it before they need to use it
  • Structured Q&A forums — create formal opportunities for concerns to be heard
  • Resistance mapping — identify specific individuals or teams who remain sceptical and engage them directly
  • Manager coaching — ensure people managers can handle the questions they'll receive at go-live
Month 1
Activation

Launch with support, not just communications

The final month is not when change management starts — it is when everything that has been built over five months becomes visible. The launch should feel like a supported transition, not an abrupt switch.

  • Hypercare support — dedicated help available in the first two weeks post-go-live
  • Daily check-ins with team leads in the first week
  • Fast-track issue resolution — a clear route for people to flag problems and get answers quickly
  • Celebrate early wins — share the first success stories immediately
  • Track adoption, not just usage — are people actually doing things differently?
Months 1–3
Post go-live

Sustain and embed — the phase most programmes skip

The first 90 days after go-live are when adoption habits form or fail to form. This phase is frequently deprioritised once the launch milestone has passed. It is also when the investment in the previous five months pays off — or doesn't.

  • Monthly outcome reviews — is adoption translating into measurable business improvement?
  • Continued manager support — people managers remain the single biggest predictor of adoption
  • Second-wave training — for those who weren't ready at go-live, or who need more depth
  • Share results widely — demonstrate that the change has delivered, with specific numbers

The manager problem — and why it must be addressed early

If there is one single thing that derails late change management more than anything else, it is unprepared frontline managers.

Frontline managers are the primary channel through which most people experience organisational change. When someone has a concern about a new system, they ask their manager. When someone is struggling to adapt, they look to their manager for signals about whether that's acceptable. When someone is wondering whether this change is really serious, their manager's behaviour is the most reliable indicator they have.

Preparing managers for these conversations requires time. They need to understand the change in enough depth to answer questions confidently. They need to know what to say when someone pushes back. They need to see clearly why the change matters — not just as a programme objective, but in terms of what it means for their team's work.

This cannot happen in a briefing two weeks before launch. It requires structured engagement starting at month five or six — early enough for managers to develop their own perspective, ask their own questions, and arrive at a position of genuine confidence rather than rehearsed talking points.

What to do if you've already started late

If you're reading this with a go-live in six weeks rather than six months, the honest answer is that you cannot do everything a six-month programme would achieve. But you can triage.

Focus ruthlessly on three things: manager briefing (do this immediately, even if it's imperfect), a structured forum for concerns to be heard (resistance that has nowhere to go becomes harder resistance), and a hypercare plan for the first two weeks post-launch. Everything else is secondary.

And when the immediate pressure has passed, invest in the 90-day post-go-live phase that most programmes skip. It won't recover everything that was lost by starting late — but it is the period in which habits either form or don't, and it is still possible to intervene.

Key takeaways

Start change management when the technology decision is made, not when deployment is imminent. The six-month framework has four phases: foundation (awareness and stakeholder mapping), preparation (capability building and concern management), activation (supported launch), and post-go-live embedding. Frontline managers are the single most important lever — invest in them first and earliest.

Planning an AI rollout?

Our AI Adoption Roadmap includes a built-in change management plan — structured by phase, tailored to your organisation's size and readiness.

Get your AI adoption roadmap
← The four behaviours of successful AI adoption All insights →