From Managing Services to Managing Intelligence: Rethinking ITIL in the Age of AI

# AI

# ITIL

# Service Management

How ITIL principles can guide human judgment and AI intelligence toward better service outcomes.

October 27, 2026

Anurag Fuloria

From Managing Services to Managing Intelligence: Rethinking ITIL in the Age of AI

For decades, service management has been largely about making work more consistent, predictable and controllable.

We created processes.

We defined roles.

We established controls.

We measured performance.

We automated repetitive activities.

We integrated tools.

And ITIL provided a common language and a set of principles to help organizations create, deliver and continually improve services.

But something fundamental is changing.

Artificial intelligence is no longer simply automating tasks within existing processes. Increasingly, AI can interpret context, identify patterns, generate knowledge, recommend actions, make decisions within defined boundaries, coordinate workflows and, through AI agents, potentially execute actions across multiple systems.

This raises a bigger question for the ITIL community:

If intelligence itself becomes embedded in service delivery, are we still managing services in the same way—or do we need to rethink what it means to manage?

From automation to intelligence

The first wave of IT transformation focused heavily on automation.

A human performed a task.

Technology automated the task.

The objective was greater speed, consistency and efficiency.

AI changes the equation.

The question is no longer simply:

“How can we automate this activity?”

It increasingly becomes:

“What should happen next, why should it happen, and who—or what—should decide?”

Consider a major incident.

Traditional automation might correlate alerts, open an incident, notify the right teams and execute predefined remediation steps.

An AI-enabled environment could go further.

It could correlate signals across infrastructure, applications, user experience, recent changes, historical incidents and business context; infer a likely service impact; recommend or initiate remediation; assess whether the remediation worked; and continuously update the knowledge base.

The value is no longer only in automating individual tasks.

The value is in orchestrating intelligence across the service lifecycle.

And that distinction matters.

The emerging shift: from process execution to decision orchestration

ITIL practices have traditionally helped organizations establish repeatable ways of working.

AI introduces another dimension: decision-making at scale.

This does not make processes irrelevant. Quite the opposite.

As AI becomes more capable, organizations need clearer principles around:

what decisions can be delegated to AI;

what decisions require human judgment;

what evidence AI should use;

what boundaries must exist;

when escalation is required;

how decisions are recorded and explained;

who remains accountable for the outcome.

The future therefore may not be about replacing ITIL practices with AI.

It may be about using ITIL principles to orchestrate human and machine intelligence within those practices.

That is a fundamentally different proposition.

From Human-in-the-Loop to Human-with-the-Loop

Much of the early AI conversation focused on the concept of the human-in-the-loop.

AI recommends.

Human reviews.

Human approves.

AI executes.

This remains valuable for high-risk decisions.

But as AI capabilities mature, another model is emerging:

Human-with-the-loop.

Here, AI can continuously observe, analyse, recommend and potentially act within predefined boundaries, while humans provide judgment, exception handling, context and accountability.

The human is no longer required to participate in every transaction.

Instead, the human becomes responsible for defining the conditions under which AI can operate and intervening when those conditions are breached.

This changes the nature of several ITIL capabilities.

Incident management may involve AI agents identifying and resolving known patterns before a human ever sees an incident.

Problem management may evolve from analysing historical incidents to continuously identifying emerging patterns and systemic risks.

Change enablement may increasingly involve AI assessing change risk using a much broader context than predefined approval checklists.

Knowledge management may move from maintaining static repositories toward continuously generated, validated and contextual knowledge.

Service level management may evolve from reporting whether targets were achieved toward understanding why experience and outcomes changed.

The practices remain.

But the way work flows through them changes.

The service ecosystem is becoming more intelligent

There is another important shift.

Services are rarely delivered by a single team or technology anymore.

A business service may depend on:

People + Processes + Platforms + Data + AI Agents + Suppliers + APIs + Customers + Partners

Each element can generate signals, decisions and actions.

This creates an increasingly interconnected service ecosystem.

In such an environment, optimizing one component independently may not optimize the service.

A highly efficient service desk does not compensate for a poor customer experience.

A perfectly automated infrastructure process does not necessarily create business value.

A high-performing application does not guarantee a successful business outcome.

This is where the broader ITIL concept of value co-creation becomes increasingly relevant.

The unit of optimization may gradually shift from the individual process to the end-to-end service outcome.

Intelligence needs governance

There is, however, a critical caveat.

More intelligence does not automatically mean better outcomes.

An AI system can process enormous amounts of information and still produce an inappropriate decision if the context, objective, data or boundaries are wrong.

This makes governance increasingly important.

Organizations will need to think beyond:

“Can AI perform this task?”

and ask:

“Should AI perform this task, under what conditions, using what evidence, with what controls and with whom accountable?”

This is where established service management principles become particularly valuable.

Governance, risk management, continual improvement, measurement, stakeholder collaboration and clear accountability do not become less important in an AI-enabled enterprise.

They become more important.

AI may accelerate execution.

ITIL principles can help provide the guardrails.

From KPIs to outcome intelligence

Another transformation may occur in measurement.

Traditional ITSM environments often rely heavily on operational indicators:

incident volumes;

resolution times;

availability;

change success rates;

SLA compliance;

backlog;

automation rates.

These remain useful.

But an AI-enabled service ecosystem can generate significantly richer signals.

We can increasingly connect operational telemetry with user experience, business context, service dependencies and behavioural patterns.

This creates an opportunity to move from simply measuring **activity and performance** toward understanding outcomes and experience.

The question changes from:

“Did we resolve the incident within SLA?”

to:

“What was the actual impact on the customer, employee or business—and how can we prevent the experience from recurring?”

That is a much richer definition of service management.

Continual improvement becomes continuous learning

AI also changes the cadence of improvement.

Traditional continual improvement often depends on periodic reviews, retrospectives, trend analysis and improvement initiatives.

AI can potentially make this much more continuous.

Every interaction can become a learning signal.

Every incident can provide context.

Every change can contribute to risk intelligence.

Every customer interaction can reveal experience patterns.

Every automation can generate evidence about whether the intended outcome was achieved.

The opportunity is to create a feedback loop:

Sense → Understand → Decide → Act → Learn → Adapt

This moves continual improvement closer to a continuously learning service ecosystem.

But there is an important distinction:

Continuous learning does not mean uncontrolled learning.

The organization still needs intentional objectives, quality controls, governance, validation and human accountability.

What happens to the ITIL professional?

Perhaps the most interesting question is not what AI will do to ITIL.

It is what AI will do to the ITIL professional.

If AI increasingly handles routine analysis, documentation, classification, correlation and execution, the value of the professional may shift toward capabilities that are harder to automate:

Context.

Judgment.

Systems thinking.

Stakeholder understanding.

Ethical reasoning.

Risk management.

Orchestration.

Outcome design.

The ITIL professional of the future may spend less time managing individual transactions and more time designing the environment in which intelligent systems can operate safely and effectively.

This could represent an important evolution:

From process manager → to service orchestrator.

From workflow owner → to outcome designer.

From operational controller → to intelligence governor.

A new question for the ITIL community

The conversation around AI and ITIL often begins with:

“How can AI improve ITIL?”

Perhaps we should ask a more fundamental question:

“How should ITIL evolve when intelligence becomes abundant?”

Because the challenge of the future may not be a lack of intelligence.

It may be an abundance of it.

When AI agents can analyse millions of signals, generate recommendations, execute workflows and interact across systems, the differentiator may no longer be access to intelligence.

The differentiator may be the organization's ability to orchestrate intelligence toward meaningful outcomes.

That is where ITIL has an opportunity to evolve.

Not by becoming an AI framework.

Not by replacing human judgment with machines.

And not by turning every ITIL practice into an AI use case.

Instead, by providing principles that help organizations design services where people and intelligent systems work together, within appropriate boundaries, to create and continually improve value.

From managing services to managing intelligence

The future of service management may therefore be less about managing individual processes, tools or technologies—and more about managing the flow of intelligence across the service ecosystem.

AI can sense.

AI can analyse.

AI can predict.

AI can recommend.

AI can act.

But organizations still need to decide:

What matters?

What should be optimized?

What decisions can be delegated?

What must remain human?

What risks are acceptable?

And, ultimately, what outcome are we trying to create?