Key takeaways

• AI success depends on fixing foundational data and asset issues first
• Enterprise architecture is shifting from systems of records and systems of engagement to systems of agency and action
• Most employees are already ahead of their AI Governance
• Poor oversight can lead to agent sprawl and inconsistent decision-making
• Regulated industries may be among the biggest beneficiaries of stronger AI governance

AI conversations often start in familiar places: chat interfaces, automation and productivity gains. Those topics matter, but they are only part of the picture. The more meaningful shift is happening deeper inside the enterprise, where architecture, workflows and governance are being reshaped to support a new generation of AI-driven action.

That is why governance has become such an important differentiator. Organizations are under pressure to move quickly with AI, but speed alone will not create lasting value. The real advantage comes from building the right foundations and creating a model where AI can operate with visibility, accountability and trust.

This is also why platforms such as ServiceNow are becoming more important to the AI discussion. For many enterprises, ServiceNow powers the workflows, service operations and data relationships that connect the business. That makes it a natural place to think not just about automation, but about how AI can be governed as it moves from assistance to action.

The move from records to agency

Enterprise architecture will evolve through three layers. The first two layers have evolved over decades. 

1. The first is the system of record

This is where the source of truth lives: ERP systems, CMDBs, ticketing history, asset data and the relationships that define how the enterprise works. Imagine this as the “basement” of your house. 

2. The second is the system of engagement

This is where users interact with those systems through portals, personas, collaboration tools and workflow orchestration. The “shiny” front door and welcoming porch of your house. 

3. The third is the system of agency

This is now emerging as a major new layer in the enterprise. It is where AI agents can make decisions, take actions, respond to feedback and improve over time.

This third layer changes the conversation. Enterprises are no longer just storing information and managing interactions. They are beginning to create environments where AI can act, which creates significant opportunity, but also raises a more urgent need for governance. Imagine building a new floor in your house, with elevators to the other floors, even while your kids and partner want the best experience living there.

The three challenges enterprises are facing

Across customer conversations, we see three common challenges. 

1. The first is still about getting the basics right

Many organizations continue to struggle with poor CMDB quality, incomplete asset inventories and disconnected systems. These are real operational issues that limit progress. If the underlying data is weak, it becomes much harder to scale AI with confidence.

This is also where ServiceNow often becomes highly relevant. If CMDB data is incomplete, asset visibility is weak or systems are poorly connected, organizations will struggle to build reliable AI services on top. Before enterprises can scale Agentic AI, they need confidence in the operational data and workflows that platforms such as ServiceNow are designed to bring together. 

2. The second challenge is the experience layer

Many businesses have built polished persona-based front ends and modern workflows, but the experience does not always connect effectively to the systems and data behind it. A strong interface alone is not enough. 

3. The third challenge is knowing how to start with AI

There is strong momentum in the market and a genuine fear of missing out. Some organizations are experimenting in IT and then expanding into areas such as HR. Others are asking a broader question around how to establish the right governance model before scaling further.

These challenges appear in different combinations depending on the organization, but they all point to the same conclusion: successful AI adoption requires more than deploying new tools.

Why governance must span all three layers

Many organizations are tempted to place an AI layer on top of their existing environment and expect value to follow. In my experience, that approach often leads to a difficult journey.

The organizations that create real value from AI are the ones that rethink the relationship between the system of record, the system of engagement and the system of agency. Governance must span all three layers. It must connect reliable source data, well-structured workflows and clearly governed agents.

That is the three-layer governance model enterprises need to adopt. Governance should not sit off to the side as a separate exercise. It must be embedded into the way AI is introduced and scaled across the business.

Why a control layer matters

As AI agents become part of the enterprise, leaders need visibility into what those agents are doing, where they are operating and what level of autonomy they have been given.

An agent may identify vulnerabilities, flag risky IT changes or even initiate remediation. That can be powerful. But enterprises still need clarity on where autonomous action is acceptable and where humans need to stay in the loop.

This is why a control layer is so important. The ServiceNow AI Control Tower is designed to provide that visibility across the enterprise, giving leaders a clearer view of where AI agents are operating, what policies apply and how value is being created. Just as importantly, it helps define the boundaries between autonomous decision-making and human oversight, so organizations can scale AI with more confidence and consistency.

That matters because governance should not be treated as a brake on innovation. Used properly, a capability such as ServiceNow AI Control Tower can help organizations move faster by making AI activity more visible, auditable and aligned to enterprise policy. In that sense, governance is not a barrier to scale. It is one of the key enablers of it.

When it is done well, governance gives organizations the confidence to move faster because they can see how AI is being used and where controls are in place.

The bigger risk is agent sprawl

Most organizations understand that governance matters. The risk is that they address it too late or too narrowly.

As agents gain more context about the enterprise, they also gain more influence. They understand systems, decision points and exceptions. Without the right oversight, that can lead to agent sprawl, where multiple agents operate across the business without enough consistency, transparency or control.

Governance will follow a path like cybersecurity. A decade ago, security was often treated as an afterthought. Today, it is embedded into every stage of application development and infrastructure operations. Governance needs to become just as integrated into the AI life cycle.

Why this matters so much in regulated industries

This issue is especially important in sectors such as financial services, insurance and healthcare, where governance has always been central to business operations.

In these industries, stronger governance may accelerate AI adoption. When leaders have visibility across platforms, policies and enterprise activity, they are in a better position to adopt AI with confidence. In other words, governance is not just about reducing risk. It can also be a catalyst for faster and broader adoption in high-trust environments.

The practical value of getting the basics right

There is a strong strategic conversation to be had around AI operating models and enterprise blueprints. That matters. But in many cases, the greatest value still comes from solving foundational problems first.

This is also where the partnership between HCLTech and ServiceNow becomes foundational. The value is not only in shaping AI strategy, but in helping customers improve CMDB quality, strengthen asset visibility and connect the systems that matter most. In many organizations, that foundational work is what makes later AI adoption more effective.

Getting CMDB data in shape, improving asset visibility and strengthening integrations across systems remain critical priorities. These are long-standing enterprise issues, and they do not disappear in the AI era. In many ways, they become even more important.

That is why the path to AI value often begins with a very practical step: fix the basics, then scale with governance in place.

Build the governance muscle now

My advice to CIOs and enterprise leaders is simple: don’t wait for AI to settle before focusing on governance.

Build that muscle now. Put a control layer in place. Create visibility across systems and models. Define policies, audit decisions and link AI initiatives back to enterprise architecture and strategic priorities.

For organizations already using ServiceNow, that can mean starting with a stronger control layer and using it to create visibility across models, policies, workflows and enterprise systems. The goal is not simply to add more AI. It is to ensure AI is introduced in a way that is governed, measurable and tied to business priorities.

The organizations that lead in the AI era will not simply be the ones deploying the most agents. They will be the ones that can govern AI effectively across records, engagement and action, turning experimentation into sustainable business value.

FAQs

Why is governance important in the AI era?
AI is moving beyond simple assistance into decision-making and action. That requires clear policies, accountability and oversight.

What are the three layers of enterprise architecture for AI?
They are the system of record, the system of engagement and the system of agency.

What is agent sprawl?
Agent sprawl happens when multiple AI agents are deployed across the enterprise without enough visibility, coordination or control.

Why are data foundations still so important for AI?
AI depends on reliable data, accurate asset records and strong integrations. Without those foundations, it is difficult to scale AI successfully.

How can regulated industries adopt AI with confidence?
By embedding governance into every stage of the AI life cycle, using strong controls and maintaining visibility across systems and decisions.