Executive Summary
CDO Magazine's April 2026 Trend Report, AI and Data Governance in the Enterprise, is an important market signal for every executive trying to move AI beyond experimentation. The report brings together an expert letter from Doug Llewellyn, CEO of Data Society Group, research findings by Dr. Tuhin Chattopadhyay, Professor of AI and Analytics at JAGSoM, and practical contributions from enterprise leaders across governance, privacy, security, automation and agentic AI.
The report matters because it does not treat AI readiness as a technology issue alone. It frames AI scale as a governance, trust and operating model challenge. The headline finding is striking: AI adoption is widespread, but only a small minority of organizations have reached Scaled maturity. The report identifies four recurring structural gaps preventing scale: limited data visibility, reactive risk management, weak accountability mechanisms and insufficient ecosystem oversight.
For Data Tiles, this validates the direction we have been championing for some time. AI readiness cannot be achieved through models, infrastructure or dashboards alone. It depends on trusted, governed, fit-for-purpose data reaching the point where decisions are actually made. That is the shift from data-driven to decision-driven. It is also why business-owned data products, active governance and governed AI agents are becoming central to the next phase of enterprise data strategy.
Latttice and Lenz were built for this market reality. Latttice enables business teams to create trusted, governed data products because they understand the business context, decision consequence and operational urgency. Lenz builds on those governed data products so AI agents can operate on trusted foundations rather than disconnected, undocumented or uncontrolled data.
Why This Report Matters
I want to begin by acknowledging CDO Magazine, Data Society Group, Doug Llewellyn and Dr. Tuhin Chattopadhyay for putting an important conversation into a clear enterprise frame.
There are many reports on AI. There are fewer reports that address the practical reason AI struggles to scale inside large organizations. This report does that well. It moves past excitement around AI adoption and asks a more useful question: what separates organizations experimenting with AI from those able to scale AI safely, consistently and with confidence?
Doug Llewellyn's opening letter is worth pausing on. He describes governance as the framework that helps organizations stay “safe while we move fast.” That phrase matters. Too often governance is discussed as bureaucracy, compliance drag or the thing that slows innovation down. Doug's point is the opposite. Governance is what allows speed to be sustained. Without it, the risks compound faster than the benefits.
Dr. Tuhin Chattopadhyay's research findings give structure to that conversation. The report assesses AI readiness across governance capability, data and architecture readiness, risk and compliance readiness and organizational confidence. It then classifies organizations into Foundational, Emerging, Operational and Scaled maturity stages. That structure is useful because it shows AI maturity is not a single technology curve. It is a composite operating capability.
That is the point executive teams need to absorb. AI readiness is not achieved because an organization adopts AI. AI readiness is achieved when the organization can repeatedly apply trusted data, governance, accountability and risk controls to the decisions and workflows AI is meant to support.
The Report's Core Finding: Adoption Is No Longer The Problem
The first major signal from the report is that enterprise AI has moved beyond experimentation. The majority of organizations are now in the Emerging or Operational stages of maturity. That tells us the market has crossed an important threshold. AI is no longer just being explored by innovation teams or isolated business units. It is moving into enterprise programs, operating models and strategic roadmaps.
But the report also shows that only a small minority of organizations have reached Scaled maturity. That gap between adoption and scale is the story.
Many organizations confuse activity with maturity. A company can have multiple AI pilots, a strategy deck, a technology partner, a data platform and a governance framework and still not be ready to scale AI. The question is not whether AI exists inside the enterprise. The question is whether AI can be trusted, governed and repeated across business-critical decisions.
That is where many organizations are now stuck. They have invested in the technology, but the operating model has not caught up. They have governance policies, but not always governance activation. They have data assets, but not always trusted data products. They have AI ambition, but not always the business ownership, traceability and accountability required to turn AI into reliable execution.
The market signal is clear: AI adoption is becoming common, but AI scale is becoming the differentiator.
The Four Gaps Holding AI Back
The report identifies four structural gaps that consistently limit AI scale: limited data visibility, reactive risk management, weak accountability mechanisms and insufficient ecosystem oversight. Each one is important on its own. Together, they describe why so many organizations remain trapped between pilot success and enterprise impact.
Limited data visibility means organizations may know data exists, but cannot always trace where it came from, how it moved, how it changed or whether it remains fit for purpose. This is a major issue for AI because AI systems and agents do not simply need access to data. They need context. They need lineage. They need to know whether the data can be trusted for the decision being made.
Reactive risk management means organizations are still assessing risk too late in the process. Risk governance has to move earlier. It has to be embedded before data is consumed, before AI is deployed and before decisions are automated. If risk is only reviewed after the fact, the organization is not governing AI. It is auditing the consequences of AI.
Weak accountability mechanisms are also a serious concern. AI introduces new questions of ownership. Who owns the data product? Who owns the decision logic? Who owns the policy? Who owns the outcome? Who is accountable when an AI agent acts on information that was incomplete, outdated or used out of context? Without clear ownership and enforceable controls, AI scale creates organizational ambiguity.
The fourth gap, ecosystem oversight, may become one of the most important. AI is not being built entirely inside the enterprise. It depends on vendors, platforms, models, data providers, partners and third-party services. If governance does not extend across that ecosystem, organizations may believe they have control when in reality they only have partial visibility.
The report is clear that trusted AI will not be built through isolated controls. It requires integrated governance across data, risk, operations, legal accountability and ecosystem oversight. I would add one more point: integrated governance only creates value if it reaches the point of decision.
Beyond The Benchmark: What The Practitioner Voices Tell Us
The report is particularly useful because the research is reinforced by practitioner perspectives.
Lana DeMaria, Head of Data Governance and Privacy at Alaska Air Group, writes about the “dark matter” of the enterprise: the unstructured information contained in emails, PDFs, messages, transcripts, contracts and other content where much of the business context lives. Her warning is important because AI will not only consume structured data. It will increasingly interact with documents, conversations, knowledge assets and operational context. If that information remains unmanaged, organizations will struggle to govern the very material AI needs to be useful.
Nick Ritter, CISO at Worldpay, makes a similarly practical point in his discussion of modern data protection. His framing around balancing governance, risk and functionality is valuable because it recognizes a reality every enterprise faces. If governance creates too much friction, people work around it. If it creates too little control, risk expands. The future is not heavy-handed governance. It is intelligent governance that protects the business while allowing work to happen.
Shivanku Misra, Global Head of AI and Automation at McKesson, brings the discussion directly into agentic AI. His chapter notes that agentic AI turns language models into “software collaborators” by giving them structured access to data, workflows and policies. That is exactly why governance becomes more important, not less important, as AI becomes more capable. The question is no longer whether organizations should use AI. It is which processes AI should touch and how safely those processes can be governed.
John Tucker, Global Director of Data and AI Governance at McDonald's, makes one of the clearest points in the report. He argues that manual controls can no longer keep up with the pace of modern data ecosystems and that automation transforms governance into a “strategic enabler.” His supplier data example is important because it shows governance moving from periodic review into operational execution. Data quality, metadata, lineage, access, privacy and stewardship become embedded into the workflow, not checked after the damage is done.
Taken together, these practitioner views point to the same conclusion as the benchmark research. Governance is moving closer to operations. AI is moving closer to decisions. Ownership is moving closer to the business. The organizations that connect those shifts will be the ones that scale.
The Governance Imbalance
One of the strongest insights in the report is that advanced AI governance capabilities are emerging before the full governance foundation is complete. Organizations are adopting real-time policy enforcement and AI-specific controls, but foundational capabilities such as data mapping, proactive risk assessment, contractual accountability and third-party oversight remain less mature.
That is a governance imbalance.
It means enterprises are building advanced control mechanisms on top of incomplete foundations. On paper, that may look like progress. In practice, it creates fragility. You can automate a policy, but if the underlying data lacks context, lineage, ownership or trust, the automation does not solve the problem. It may simply make the problem move faster.
AI amplifies whatever operating model sits beneath it. If the organization has fragmented data, AI will amplify fragmentation. If governance is passive, AI will expose that passivity. If ownership is unclear, AI will create more accountability gaps. If data is not trusted by the business, AI will not magically make it trustworthy.
The report's finding should be a warning to executive teams. Do not mistake AI controls for AI readiness. Controls matter, but they must sit on top of trusted, governed, business-owned data foundations.
Confidence Is A Consequence, Not A Starting Point
Another important finding is the relationship between maturity and confidence. The report argues that confidence increases as governance maturity increases. That may sound obvious, but it is actually a critical point.
Many organizations are waiting to feel confident before they scale AI. The more mature organizations are doing the opposite. They are building the governed, repeatable systems that create confidence.
Confidence does not come from enthusiasm. It comes from evidence. It comes from knowing where data came from, who owns it, whether it is fit for purpose, what policies apply, who can access it, where it has been used and whether it can be trusted in the context of a decision.
That is why confidence cannot be separated from governance. It is the outcome of trust being operationalized.
The Hybrid Operating Model Is Winning
The report also highlights the rise of hybrid data cultures, where centralized governance is combined with decentralized execution. This is one of the most important signals for the future of enterprise data and AI.
The old debate was often framed as centralization versus decentralization. Centralized models created consistency and control, but often moved too slowly for the business. Decentralized models created speed and domain ownership, but often introduced inconsistency, duplication and risk. The market is now moving toward a more pragmatic model: central governance standards, decentralized business execution and shared accountability.
That is exactly where business-owned data products fit.
The people closest to the decision understand the business context. They understand the consequence of being wrong. They understand the urgency. They understand the difference between data that is technically correct and data that is actually fit for purpose.
But business ownership cannot mean uncontrolled data creation. It has to be governed by design. That is the role of active governance. It allows business teams to move faster without moving outside the boundaries of trust, policy and accountability.
Cameron Price: Why This Moment Matters
For me, this report confirms what I have believed for a long time. The future of enterprise data will not be defined by who has the most data. It will be defined by who can be trusted at the point of decision.
I have spent much of my career around enterprise data, analytics, governance and cloud transformation. I have seen organizations invest heavily in platforms, warehouses, lakes, catalogs, dashboards and governance programs. Many of those investments were necessary. But they did not always solve the problem the business actually cared about: can I get trusted data when I need to make a decision?
That is the gap Data Tiles was created to close.
The market spent years trying to make organizations more data-driven. I believe the next stage is decision-driven. Data only creates value when it changes a decision, improves an action or reduces uncertainty. If a data product does not change a decision, it is not really a product. It is another report.
That belief shaped Latttice. We built Latttice because business teams need a way to create trusted, governed data products themselves. Not because data teams are less important. They are essential. But the business understands the context, the question, the urgency and the consequence. The role of technology should be to help them create trusted products safely, not force every decision through long technical delivery cycles.
It also shaped Lenz. AI agents cannot safely operate on unknown, disconnected or undocumented data. They need trusted business-owned data products, active governance, access controls, lineage, context and accountability. If an AI agent is going to support a decision, the foundation beneath it must already be trusted.
That is why Latttice and Lenz are connected by design. Latttice activates trusted data products. Lenz activates governed AI agents on top of those products. Together, they support the operating model I believe the market is now moving toward: active governance at the point of decision.
What We Are Seeing In North America
“In North America, we are seeing the conversation move quickly from AI ambition to AI accountability. A year ago, many organizations were still asking how to start with AI. Today, most have already started — they have pilots, teams experimenting, vendors in the conversation and executives asking for momentum. The more serious conversations now are about trust. Where did this data come from? Who owns it? Can the business trust it? Can an AI agent safely use it? Those questions are no longer technical details — they are becoming executive questions. North American organizations want speed, but they also want control. That is why business-owned data products are resonating. They give organizations a practical way to connect business context with governed data while ensuring policies, permissions and trust are built in.”
What We Are Seeing In the UK & EU
“In the UK and Europe, governance activation is becoming the defining theme of the AI readiness conversation. Many organizations here already have mature governance frameworks — councils, policies, stewardship models, regulatory reporting structures — but most were designed for an era when governance happened around data, not inside the decisions and AI agents using it. Leaders are increasingly candid that a well-documented policy in a portal is no longer enough when an AI agent is acting on customer data in real time. What we are seeing across UK and EU organizations is a shift from governance as oversight to governance as activation: embedding ownership, lineage, permissions and business meaning directly into the data products that decisions and agents consume. The organizations making progress are turning governance into something the business experiences as enabling, not something the business experiences as friction.”
What We Are Seeing In APJ
“Across APJ, we are seeing the same groundswell with regional nuance. Organizations are highly motivated to adopt AI — the ambition, the investment and the executive attention are all there. But many are discovering that pilots are much easier than scale. The constraint is not enthusiasm. It is trust. APJ is a diverse region: markets are different, regulations vary, transformation maturity differs by country, sector and organization. One pattern is consistent — organizations need governed, fit-for-purpose data that business teams can actually use in the decisions they make every day. For AI agents, that requirement becomes even more important. What we are seeing across the region is a move toward practical AI readiness: leaders bringing governance closer to the business, empowering teams without creating chaos, and respecting regulatory and operational complexity.”
Partner & Customer Success Perspective
“From a partner and customer success perspective, what stands out in this report is the gap between AI ambition and operational confidence.
Many organizations are not short of ideas. They have use cases, executive interest and a clear desire to move faster with AI. The difficulty comes when those ideas need to become repeatable, governed and trusted in day-to-day business environments.
That is where the conversation changes.
Customers are not just asking how to deploy AI. They are asking how to make AI usable, trusted and safe for the people who need to rely on it. They want to know how business teams can create data products without creating new risk. They want to know how governance can be embedded without slowing everything down. They want to know how partners can help them move from pilot activity to practical adoption.
This is why the report's focus on integrated governance is so important. Governance cannot remain something people only encounter at the end of a project. It has to become part of the way work happens. It has to support the business user, guide the partner ecosystem and give customers confidence that what they are building can scale.
What I see with customers is that AI readiness becomes real when the business can participate. The people closest to the process understand what the data should mean, where the risk sits and what decision needs to be improved. When those teams can build trusted, governed data products with the right guardrails in place, adoption becomes far more practical.
The same is true for partners. Partners are critical because they help organizations translate strategy into delivery. But for partners to accelerate AI readiness responsibly, they need a clear operating model: trusted data products, active governance, clear ownership and repeatable patterns that can be applied across functions, regions and industries.
That is why Latttice and Lenz matter in customer and partner conversations. Latttice gives business teams a governed way to create and use trusted data products. Lenz gives organizations a path to governed AI agents built on those trusted foundations. Together, they help turn AI readiness from an executive aspiration into something customers and partners can actually implement.
For me, the strongest signal is this: AI readiness will not be achieved by technology teams alone. It will be achieved when business users, data teams, governance leaders and partners can work from the same trusted foundation.”
Lili Marsh
Partner & Customer Success
What Data Tiles Is Already Doing In This Space
The report predicts the need for integrated governance. Data Tiles was built around that principle.
Latttice is our Data Product Workbench. It enables business teams to create trusted, governed, AI-ready data products without code. The purpose is not to bypass governance. The purpose is to activate it. Governance, lineage, access controls and trust signals are built into the way data products are created and consumed.
Lenz builds on that foundation. It enables governed AI agents to operate on trusted data products rather than disconnected or ungoverned sources. AI agents will only be as reliable as the data, governance and context beneath them. If the foundation is weak, the agent is weak. If the foundation is trusted, governed and business-owned, AI has a much stronger path to enterprise adoption.
This is why we see Latttice and Lenz as part of the same operating model. Latttice creates and governs the trusted data product layer. Lenz uses that layer to support governed AI agents. Together, they bring trusted data and active governance to the point of decision.
The Market Signal Worth Watching
The most important signal from the report is not that organizations are adopting AI. We already know that. The signal is that AI scale is being constrained by governance fragmentation.
AI readiness is not just a technology checklist. It is not just model selection, infrastructure design or cloud investment. It is the ability to repeatedly connect trusted data, governance, business ownership, accountability and decision-making.
The organizations that succeed will be those that treat governance as an operating capability, not a documentation exercise. They will move governance closer to the business. They will create data products around decisions, not just datasets around systems. They will build AI agents on trusted foundations, not uncontrolled information flows. They will measure decision impact, not just data availability.
That is the market we believe is emerging.
It is a market where trust becomes operational infrastructure.
It is a market where governance moves from passive oversight to active enablement.
It is a market where business-owned data products become the bridge between enterprise data and AI execution.
And it is a market where AI readiness is achieved at the point of decision.
As the market shifts from AI adoption to AI scale, leaders should ask themselves:
- Can our AI initiatives access trusted, governed data at the point of decision?
- Is governance active in the flow of work, or still a documentation exercise?
- Do business teams own the data products they depend on?
- Can we trace every AI-driven decision back to a trusted, accountable source?
- Are we investing in AI controls, or in the foundations that make AI trustworthy?
- Are we building for AI experimentation, or for AI at scale?
The organizations that can answer these questions confidently will be best positioned to move beyond AI experimentation and into repeatable, governed business outcomes.
Because the real signal within CDO Magazine's 2026 Trend Report is not about AI itself.
It is about trust.
And in the AI era, trust may become the most valuable asset an organization possesses.
Closing Perspective
CDO Magazine's report is valuable because it gives executives a clear view of where the market really is. AI adoption is no longer the differentiator. AI scale is. And AI scale depends on integrated governance, trusted data, business ownership and accountability.
For Data Tiles, this is not a new direction. It is the reason we built Latttice and Lenz.
We believe the next phase of enterprise advantage will belong to organizations that can act with confidence when decisions need to be made. Not because they have more data. Not because they have more dashboards. Not because they have more AI pilots.
Because they have trusted, governed, fit-for-purpose data at the point of decision.
That is the foundation for AI readiness.
That is the future we are building toward.