Organisational Change and Adoption
New ways of working, tools, and AI succeed when people understand why, take part in shaping them, learn by doing, and see that leaders keep their word; plan for adoption as the hard part of any change, start where you are, and measure whether people use the change, not whether it shipped.
Why this matters
Most failed technology projects did not fail on technology. The system was delivered, the process was documented, the training was held, and six months later people were doing the work the old way, or in a workaround, or not at all. The investment bought a capability that nobody used.
Service management is full of such changes: a new ticketing platform, a move from a change advisory board to pipeline controls, a self-service portal, an AI assistant on the service desk, a shift to persistent product teams. Each changes what people do every day, who decides, what skills matter, and sometimes whether a job exists. People notice. They ask what it means for them, whether they will be good at it, and whether leaders are being straight with them. The answers decide whether the change takes hold.
Change management as a discipline concerns this human side. Wikipedia describes it as the approach to transitioning individuals, teams, and organisations to a desired future state. Much of what is written about it is too heavy for a team to use, and some is too glib. This topic gives a practical account: how to understand who is affected, how to involve them, how to support learning, how to handle resistance as information, and how to treat the specific fears that automation and AI raise. It is different from "change enablement" in ITIL, which concerns authorising changes to services (see Topic 3.4 — Change Enablement). It connects to agile management and learning (Topic 2.10 — Agile Management and Learning), the AI operating model (Topic 5.8 — AI Operating Model and Skills), and learning paths (Topic 8.5 — Learning Paths and Certification).
Core concepts
Organisational change. A shift in how an organisation works: structure, process, technology, roles, or culture. Wikipedia describes change management and organisational behaviour. Changes vary in size, from a new form to a new operating model, and the effort for adoption scales with how much people's daily work changes.
Adoption. The extent to which people actually use the change as intended. Adoption is not training attendance or go-live. It is observable: how many people use the new tool, in which tasks, and whether the old route has stopped. Treat adoption as a measurable outcome of the project.
Start where you are. ITIL 4's guiding principle applies to change: understand the current state, the strengths to keep, and what people value in the present way of working before replacing it (see Topic 1.5 — ITIL Guiding Principles). People will defend what they know works.
Why, what, and what it means for me. Three questions every affected person asks. Why is this change happening? What exactly will change? What does it mean for me: my tasks, skills, status, security, and workload? Leaders often answer the first and neglect the third.
Models of change. Several well-known frameworks describe the process. Kurt Lewin's classic model of unfreezing, changing, and refreezing, John Kotter's eight-step approach (create urgency, build a coalition, form a vision, communicate, remove obstacles, create short-term wins, build on the change, anchor in culture), and the ADKAR model (awareness, desire, knowledge, ability, reinforcement) from Prosci are widely used. Wikipedia describes Lewin, Kotter, and a range of models under change management. Use them as checklists of what to think about, not as recipes.
Stakeholders and impact. People affected differ in how much the change touches them, how much influence they hold, and their attitude. Wikipedia describes stakeholder analysis. Map who is affected and how, identify those with influence, and find out what each group needs. Include customers and suppliers where relevant.
Resistance as information. What looks like resistance often signals a real problem: a flaw in the design, a gap in training, a loss people reasonably fear, or a history of changes that were announced and abandoned. Listen before you overcome. Wikipedia discusses resistance to change as part of change management. Some resistance arises from legitimate expertise: people who know the work best often see the risks first.
Adoption curves. Everett Rogers's diffusion of innovations, which Wikipedia describes, observes that people adopt new things at different rates: a few early adopters, a large middle, and some who adopt late. Technology adoption life cycle, also described there, applies the idea to products. Work with early adopters to learn and spread, and expect the middle to need evidence from peers.
Champions and communities of practice. People inside teams who understand the change, help colleagues, and give feedback to those leading it. Communities of practice, described on Wikipedia, bring practitioners together to share what they learn. They work best when members volunteer, have time, and have a voice.
Training and learning. Courses help with concepts. Skills grow through practice on real work, with coaching and feedback. Wikipedia describes coaching. Provide time to learn, safe places to make mistakes, and support after go-live, when real questions arise. See Topic 8.5 — Learning Paths and Certification.
Communication. Regular, honest, two-way communication, using several channels, from people that staff trust. Say what is decided, what is not, and what is unknown. Silence is filled with rumour. Listen as much as you tell.
Psychological safety and trust. People adopt change when they feel safe to try, to fail, and to raise concerns. Wikipedia describes psychological safety and trust. Trust is built by keeping commitments, and it is lost when promises, such as "no jobs will be lost", are broken. See Topic 2.10 — Agile Management and Learning.
Change load and fatigue. Too many changes at once exhaust people. Wikipedia describes change fatigue. Count the total change demand on a team, sequence changes, and finish and embed before starting the next. Dropping old initiatives explicitly frees capacity.
Short wins and feedback. Early, visible benefits build support, and honest feedback tells you what to fix. Start with a pilot that helps real users, learn from it, and widen. See Topic 2.9 — Agile Service Management.
Reinforcement and anchoring. A change sticks when the old way is retired, roles and incentives support the new one, and leaders use it themselves. If the old report is still demanded, the new tool will be used as an extra step.
Automation and AI: specific concerns. People fear for their jobs, worry about being monitored, doubt the accuracy of AI, and wonder what skills will matter. Wikipedia describes technological unemployment and technostress as concerns that accompany new technology. Be honest about intentions, involve staff in deciding what to automate, retrain and redeploy where roles change, and make accountability clear: staff are not blamed for AI errors they were told to accept. See Topic 6.1 — Automation Strategy and Topic 5.8 — AI Operating Model and Skills.
Culture. The shared habits and assumptions of an organisation. Wikipedia describes organisational culture. Culture eats strategy: a change that conflicts with how things are really done will be absorbed or rejected. Work with the culture, and change it through what leaders do and reward, not through slogans.
Measuring adoption. Look at use, quality of use, time to proficiency, support demand, and how people feel. Combine system data, observation, and short surveys. See Topic 7.7 — Experience and Sentiment Analytics.
How the lenses fit. Change and adoption are the human side of all four lenses. Agile practice gives the shape: small steps, feedback, and involvement. AI and automation raise the issues of trust and jobs most sharply. Analytics shows whether the change is used and helps.
Best practices
- Be clear about the why. State the problem, the evidence, and what will be better, in plain words and with the people affected. Check that they understand and, ideally, agree.
- Map who is affected and how. List groups, what changes for each, what they stand to gain or lose, and their influence. Include customers and suppliers. Talk to each group.
- Start where people are. Learn what works today and what people value. Keep what is good. Avoid describing the present as a failure.
- Involve people in the design. Bring affected staff into choosing and shaping the change, through workshops, trials, and review. People support what they help create.
- Pilot, learn, widen. Begin with willing teams and a real problem, gather feedback, fix what you learn, and show results before extending.
- Support learning by doing. Provide time, coaching, practice on real tasks, and help after go-live. Create a community where people share tips and problems.
- Listen to resistance. Ask what concerns lie behind it, and act on the valid ones. Thank people who raise problems.
- Communicate honestly and often. Share what is decided, what is not, and when you will know more. Use several channels and two-way forums. Do not promise what you cannot guarantee.
- Manage the load. Keep a view of all changes affecting each team, sequence them, and stop or defer lower-value ones to make room.
- Measure adoption and close the old route. Track real use, retire the old way on a stated date, and align roles, reports, and rewards with the new way. Ask users what is still hard.
Questions to discuss with your team
What problem does this change solve for the people who will live with it, and do they agree it is a problem? Ask them. Expect that the problem is clearer to leaders than to staff. Follow up by gathering one or two stories from daily work that show the problem.
What do people fear losing, and what are they right to worry about? Speak plainly about status, skills, workload, and jobs. Expect some fears to be valid. Follow up by deciding what you can honestly commit to.
Who are the people others listen to, and have we talked with them? Think of respected colleagues, not only managers. Expect that some were overlooked. Follow up by arranging a conversation and asking for their view of the plan.
What other changes are hitting this team at the same time? Add them up. Expect more than anyone knew. Follow up by agreeing what to pause, finish, or sequence.
How will we know that people are using the new way, not just that it has gone live? Choose observable signs, such as share of requests through the new route. Expect that go-live was the planned measure. Follow up by agreeing adoption measures and a review date.
How will AI or automation change roles here, what have we said, and what are we prepared to promise? Ask what staff have been told. Expect rumour to fill gaps. Follow up by drafting an honest statement with leaders and checking it with staff representatives.
In practice: a service management example
A national tax agency's service desk of 120 agents is about to introduce an AI assistant that drafts replies and suggests knowledge articles. Past changes at the agency, such as a new ticketing tool, were rolled out by email and a one-hour course, and agents say they were "told, not asked". A union representative has already written to management about fears that the assistant is a first step to cutting jobs.
The head of service starts by meeting agents in small groups, asking about their work and what slows them down. Agents say they spend much time searching for articles and rewriting similar replies, and that they dislike scripts that make them sound like machines. The head tells them honestly what is intended: the assistant will draft, agents will edit and own every reply, and the agency has no plan to reduce posts as a result for at least two years; she puts this in writing, and asks the union to review the wording.
A pilot team of twelve volunteers, chosen to include skeptics, uses the assistant for six weeks. Weekly sessions collect feedback. Agents find that drafts for common questions are useful and that those for complex cases are poor and sometimes wrong. They ask for a button to flag bad drafts, which the team adds, and for a rule that using the assistant is optional for sensitive cases. The pilot's agents become champions, and a monthly community of practice shares prompts and tips. The agency also agrees that metrics of how many drafts an agent accepts will not be used to evaluate individuals.
After the wider rollout, 70 percent of agents use the assistant weekly, and time to first reply falls modestly. Agents say they feel heard. What remains hard: about a fifth of agents rarely use it, some because the drafts do not fit their queue, some from distrust; a manager who wanted to use acceptance rates in appraisals had to be told no; and the two-year commitment about posts constrains the finance team's plans. The head of service holds a quarterly review with the union and adds a training path for agents who want to move into knowledge or quality roles.
Four perspectives
Product and delivery teams
Involve the people who will use what you build, from design to rollout. Provide training material, support, and ways to give feedback. Watch how people use it, and change the product when they struggle.
Service and operations teams
Take part in shaping changes, be honest about what will and will not work, and help colleagues learn. Raise concerns early and propose alternatives.
Leadership and governance
Explain the why, make and keep honest commitments, and use the new way yourself. Fund time for learning, and manage the total load on teams. Listen to resistance as information.
Customers and users
You may be asked to change how you ask for help or use a service. Tell the provider what makes it easier or harder, and expect to be told what is changing and why.
Agile, AI, automation, and analytics
Agile
Change in small steps with early users, feedback, and visible results, and involve stakeholders throughout. Caution: frequent small changes still add up; keep track of the total load and finish before starting more.
AI
AI can draft communications, summarise feedback, and suggest training, under review. The adoption of AI itself needs honest communication about jobs, accuracy, and accountability. Caution: do not use AI to read employee sentiment covertly or to monitor individual use; it destroys trust. See Topic 5.8 — AI Operating Model and Skills.
Automation
Automate reminders, training enrolment, and adoption data collection, and involve staff in choosing what to automate. Caution: automating a task people value, without discussion, can feel like a threat; say what the time saved will be used for. See Topic 6.1 — Automation Strategy.
Analytics
Use adoption data, support questions, time to proficiency, and short pulse surveys, at team level. Caution: usage counts can be inflated by mandates and do not show whether work improved; combine with feedback, and do not rank individuals. See Topic 7.7 — Experience and Sentiment Analytics and the Service Management Metrics book.
Common failure modes
- Announce and train. People are told, not asked. Involve them from the start.
- Go-live as success. Nobody checks use. Measure adoption.
- Ignoring fears. Rumours fill the gap. Say plainly what you know and what you do not.
- Calling skeptics resistant. Valid concerns are lost. Listen and act on them.
- Change overload. Teams are swamped. Sequence and stop lower-value changes.
- Old route still alive. The new way is an extra step. Retire the old and align roles and reports.
- Broken promises. Trust collapses. Promise only what you can keep, and keep it.
Maturity model
| Dimension | Initiate | Develop | Standardise | Manage | Orchestrate |
|---|---|---|---|---|---|
| Understanding and involvement | Told | Consulted at the end | Affected groups mapped and involved from the start | Involvement quality reviewed | Co-design is the norm across the organisation |
| Learning and support | Course only | Training plus help desk | Practice on real work, coaching, and communities | Time to proficiency measured | Learning continuous and shared across teams and partners |
| Communication and trust | One-way | Regular updates | Honest, two-way, and consistent with actions | Trust and sentiment tracked | Transparent communication across partners |
| Adoption and load | Unmeasured | Some use data | Adoption measured, old route retired, load managed | Adoption results drive decisions | Change portfolio managed with capacity and outcomes |
Checklist
- We can state the problem and the benefit in plain words.
- We mapped who is affected and what changes for each group.
- Affected people helped shape the change.
- We ran a pilot, and acted on feedback before widening.
- People have time, practice, and coaching, and a community to share with.
- We listened to concerns and recorded what we changed.
- We know the total change load on each team and have sequenced it.
- We measure adoption, retire the old route, and have kept our commitments.
AI prompts
Prepare a stakeholder map.
Here is a description of the change: [paste]. Here are the groups I think are affected: [list].
For each group, list what changes for them, what they might gain or lose, questions they are likely to ask, and a question I should ask them. Mark everything that comes from my description versus general assumptions. Do not predict their attitudes.
List what you are unsure of.
Draft a plain-language explanation for staff.
The change: [describe]. The reason: [describe]. What is decided: [list]. What is not decided yet: [list]. What we commit to: [list].
Draft a message of under 250 words for affected staff that says why, what, what it means for them, what is not decided, and where to ask questions. Do not promise anything not in my list.
List what you are unsure of.
Summarize pilot feedback.
Here is feedback from a pilot, with names removed: [paste].
Group it into themes with counts, quote two short comments per theme, separate problems with the design from problems with training or support, and list questions to ask the pilot group. Do not rate individuals or infer feelings.
List what you are unsure of.
AI skills
- enterprise-change-management: use to plan the human side of significant change.
- stakeholder-analysis: use to map who is affected and what they need.
- agile-coaching: use to coach teams through changing ways of working.
- training-needs-analysis: use to define the learning people need.
- getting-feedback: use to gather and act on feedback from people affected.
Key sources
- John P. Kotter, Leading Change (1996): a widely read account of leading large-scale change.
- Everett M. Rogers, Diffusion of Innovations (5th edition, 2003): the classic account of how new ideas spread.
- William Bridges, Managing Transitions (4th edition, 2016): on the human experience of change.
- Wikipedia, "Change management" and "Diffusion of innovations": overviews of the field.
References
- Change management — Wikipedia — https://en.wikipedia.org/wiki/Change_management
- John Kotter — Wikipedia — https://en.wikipedia.org/wiki/John_Kotter
- Kurt Lewin — Wikipedia — https://en.wikipedia.org/wiki/Kurt_Lewin
- Diffusion of innovations — Wikipedia — https://en.wikipedia.org/wiki/Diffusion_of_innovations
- Everett Rogers — Wikipedia — https://en.wikipedia.org/wiki/Everett_Rogers
- Stakeholder analysis — Wikipedia — https://en.wikipedia.org/wiki/Stakeholder_analysis
- Community of practice — Wikipedia — https://en.wikipedia.org/wiki/Community_of_practice
- Coaching — Wikipedia — https://en.wikipedia.org/wiki/Coaching
- Psychological safety — Wikipedia — https://en.wikipedia.org/wiki/Psychological_safety
- Change fatigue — Wikipedia — https://en.wikipedia.org/wiki/Change_fatigue
- Organisational culture — Wikipedia — https://en.wikipedia.org/wiki/Organizational_culture
- Technostress — Wikipedia — https://en.wikipedia.org/wiki/Technostress