A practical path to the enterprise values

In our reflections on ServiceNow World Forum Sydney, we argued that the winners in enterprise AI will start boring and scale boldly. The natural next question is what that looks like in practice.

This article sets out the answer, through the three capabilities ServiceNow is assembling and the five deliberate steps that separate AI ambition from AI results.

From assistance to execution

The first wave of generative AI helped people create, search, summarise and answer questions. The shift now underway moves beyond that assistive model towards AI that can coordinate and execute work across departments and systems.

ServiceNow Otto is intended to provide a unified AI experience through which employees express what they need in natural language. Behind that interaction, Otto identifies intent, routes work, engages the appropriate workflows or agents, and progresses the request towards completion, grounded in organisational information, policies and approval chains rather than a generic model response. Instead of learning which application, form or team owns a particular service, an employee simply states the outcome they need.

That is a genuinely different employee experience. It also raises the question that should shape every AI roadmap: what happens when the organisation behind that experience is not ready?

Context is what makes data actionable

Enterprise AI has long been framed as a data problem. Increasingly, it is a context problem. An agent can have access to accurate data and still act wrongly if it does not understand how the information relates to the request, who is authorised to act, which policy applies, what has already occurred, and what outcome the organisation is trying to achieve.

ServiceNow’s Context Engine is designed to provide this operational understanding, bringing together structured and unstructured information, permissions, policies, workflow states and historical decisions so AI can act with relevance and precision. It also records the context used for each decision, supporting auditability and explainability. Raw information helps an agent answer a question. Organisational context helps it determine whether, when and how it should act.

Trust as architecture

Autonomy raises the governance bar. When AI drafts something for human review, the consequence of an error is manageable. When an agent can access systems, initiate workflows and execute decisions, organisations need visibility over which agents and AI identities are operating, what they can access, which policies apply, how their behaviour is recorded, where human review is required, and whether the deployment is producing measurable value.

AI Control Tower is positioned as the central capability for exactly this: discovering, observing, governing, securing and measuring AI systems and agents, including AI operating outside ServiceNow, with runtime monitoring, policy enforcement and mechanisms for responding when an agent operates beyond its permitted boundaries. The Veza acquisition sharpens the identity dimension. One demonstration at World Forum showed a security graph mapping every user’s access across tables and applications, cross-referenced against the organisational structure to flag access a role does not justify, with AI agents treated as identities too. As agents take on more work, that matters enormously: an agent must never become a shortcut around access controls, and excess access should be flagged before it becomes an exploitable gap.

Governance, in other words, should be established before production, not retrofitted after adoption has begun.

People remain central

None of this reduces the importance of change management. It increases it. Employees need clear guidance on responsible use, sufficient AI literacy to assess outputs critically, confidence that feedback will be heard, and visible accountability for outcomes. Leaders must clarify priorities, communicate them consistently, and bring change champions into the tent early.

The human contribution becomes more valuable, not less, as AI absorbs transactional work. Interpretation, intuition, empathy and judgement remain essential wherever context is incomplete or consequences extend beyond what a workflow can represent. The aim is not to remove people. It is to remove unnecessary friction so people can focus where their expertise has the greatest value.

Five deliberate steps

These are the five deliberate steps we recommend, drawn from what succeeded on stage at World Forum and from the transformation disciplines we apply every day:

• Begin with the outcome. Define the service, employee, customer or operational problem before selecting the AI capability.
• Redesign the work. Simplify the process and clarify ownership before accelerating it.
• Establish the foundations. Address data quality, integration, security, governance, permissions and platform performance.
• Prove value deliberately. Start with a controlled, high-volume use case and measure whether it improves the intended outcome.
• Scale with people, not around them. Build AI literacy, involve change champions, gather feedback, and adapt the operating model as capability grows.

Why sequencing decides who benefits

ServiceNow’s 2026 Enterprise AI Maturity Index found that 59 per cent of organisations have moved beyond initial agentic AI pilots, yet only 9 per cent have made meaningful progress towards autonomous, multistep workflows (source: ServiceNow, Enterprise AI Maturity Index 2026). The gap between those numbers is not a technology gap. It is a sequencing gap, and it is closable.

The organisations that benefit most from the autonomous enterprise will not simply be those that purchase the capability. They will be those that understand their services, connect their platforms, establish context, govern decisions, and bring their people through the change. As a ServiceNow Elite Partner combining platform expertise with strategy, business process, change, cyber security and delivery capability, the foundational stretch of that path, understanding services, simplifying processes, establishing governance and bringing people through change, is the work we do every day. It is where every successful AI journey starts, and for many organisations, especially in the public sector, it is exactly where the focus should be right now.

Planning where to go next? We would value the opportunity to compare notes. Talk to our team.