Change Management for Agentic AI: Preparing Your Operating Model

Agents are arriving in the enterprise whether or not anyone has decided what they mean for how the organisation actually works. Gartner expects 40% of enterprise applications to feature task-specific AI agents by the end of 2026, up from fewer than 5% in 2025. The same firm predicts that more than 40% of agentic AI projects will be cancelled by the end of 2027 — not because the models fail, but because of escalating costs, unclear business value and inadequate risk controls. Read together, those two forecasts describe one organisation: it has deployed agents far faster than it has redesigned the operating model around them.

The gap is measurable. Kyndryl’s 2026 People Readiness Report, a survey of 1,100 senior business and technology leaders, found 79% agree the speed of AI will outpace their organisation’s workforce, governance and operating models. With agentic AI, that mismatch stops being a forecast and becomes an operational problem.

Why agentic AI is a different change problem

Generative AI, for all the noise it made, largely left the operating model intact. A person still did the work; the tool made them faster. Agentic AI breaks that assumption. An agent does not assist with a task — it takes the task, executes a sequence of steps and makes decisions along the way. The change is no longer “our people work differently”. It is “some work is no longer done by a person at all, and someone must now govern what the agent does instead”.

That is an operating-model change in the strict sense. It alters who does the work, who decides, who is accountable when a decision is wrong, and how exceptions get escalated. No deployment plan answers those questions, because they are not technical questions.

The questions to settle before the agents arrive

The organisations we see struggling have almost always skipped these. The ones that succeed have written the answers down:

• Decision rights. What may an agent decide on its own? What must it recommend for a human to approve? What must it never touch? Ambiguity here is where cost and risk both escape.
• Accountability. When an agent makes a poor call, who owns the outcome — the process owner, the team, the vendor? “The system did it” is not a governance position.
• Exception handling. How quickly is a human brought back in, and do your people genuinely believe they are authorised to override the agent? Many do not, and say so only after something goes wrong.
• Role redesign. What are the humans now for? Oversight, judgement, exception handling and relationship work have to be designed and named as real roles — not left for people to infer while they wonder about their futures.
• Assurance. How will you know, six months in, that the agent is still doing what you intended?

Why the cancellations happen

Gartner’s three stated reasons for project cancellation — cost, unclear value, weak risk controls — are governance failures, not engineering ones. There is a procurement dimension too: Gartner notes widespread “agent washing”, the rebranding of existing assistants, chatbots and robotic process automation as agentic, and estimates that only around 130 of the thousands of vendors claiming agentic capability are genuine. An organisation without a clear view of the operating-model change it wants is poorly placed to judge which of those vendors it actually needs.

What the pacesetters do

Kyndryl’s research identifies a group of 9% of organisations — the “pacesetters” — who do three things the rest do not: they redesign roles around AI, they run deliberate change management so the workforce understands the new operating model and has guardrails in place, and they build workforce readiness rather than assuming it. Those organisations were 1.5 times more likely to achieve AI-related revenue growth and 1.6 times more likely to report stronger innovation in products and services. Notice that none of the three is a technology task. All three are change implementation tasks.

Preparing your operating model

This is the work CCG exists to do, and with agentic AI it has to happen before deployment rather than after:

• Set the governance and decision rights first. CCG Advisory works with Boards and executives on the oversight model, the guardrails and the accountability questions that decide whether an agentic programme is governable at all.
• Redesign the roles and the workflow. Through CCG Consult we translate an agentic ambition into changed processes, escalation paths and clearly defined human roles — the operating-model shift itself.
• Build the judgement the new roles demand. Supervising an agent is a different skill from doing the task. CCG Learn develops the confidence and capability people need to oversee, question and override these systems well.
• Measure whether it is landing. CCG Analytics tracks readiness, adoption and sentiment so you see resistance and risk building while you can still act — rather than at the cancellation review.

Agentic AI will reward the organisations that treat it as an operating-model decision rather than a purchase. The technology will keep improving on its own. Decision rights, accountability and the human roles around the agents will not — those have to be designed, and led.

If you are preparing your operating model for agentic AI, we would welcome the conversation. Get in touch with CCG.