Over the course of my career, I’ve had a front-row seat to several major technology shifts.
From IT service management and enterprise automation to cloud transformation and digital workflows, I’ve seen how new technologies create enormous excitement, attract significant investment, and promise to reshape how businesses operate.
I’ve also seen a consistent pattern emerge.
The organizations that achieve lasting business value are rarely the ones that adopt technology first. They are the ones that build the right foundation to support it.
That lesson became clear to me during my time at BMC and later at ServiceNow. The companies that successfully transformed their operations were not simply deploying new software. They were creating a shared understanding of processes, services, assets, and business outcomes across the enterprise.
Today, as I talk with CEOs, CIOs, Chief Data Officers, and AI leaders, I see a similar pattern unfolding with artificial intelligence.
Organizations are moving quickly. Many have launched copilots. Others are deploying AI agents and embedding large language models into customer service, operations, software development, and business workflows.
The enthusiasm is justified.
The challenge is that many organizations are discovering that access to AI does not automatically translate into business value.
In my view, the biggest obstacle is not the model.
It’s the data.
More specifically, it’s the gap between how enterprise data is organized and how AI systems need to understand information.
Most enterprise data was never designed for AI.
It was designed to run businesses.
It lives across ERP systems, CRM platforms, service management tools, data warehouses, data lakes, and thousands of specialized applications. Each system serves an important purpose, but each represents only part of the enterprise picture.
The result is an environment where information is fragmented across organizational boundaries, technology platforms, and business functions.
Humans are remarkably good at navigating that complexity.
AI is not.
AI needs context.
It needs to understand how customers relate to products, how assets connect to services, how policies influence decisions, and how business concepts span multiple systems. Without that context, even the most sophisticated AI can produce answers that sound convincing while missing critical business meaning.
That reality is increasingly reflected in the market.
Our strategic partner, Accenture, recently reported that 64% of organizations have moved beyond AI pilots into production deployments or enterprise-scale initiatives, yet only 7% have achieved the level of data readiness required to scale advanced AI successfully.
When I speak with enterprise leaders, those numbers ring true.
Most organizations are investing heavily in AI capabilities. Far fewer are investing at the same level in the knowledge foundation required to make AI reliable, trustworthy, and scalable.
That is why I believe the conversation around enterprise AI is entering a new phase.
The first chapter was about models.
The next chapter is about knowledge.
What I’ve Learned About Enterprise Semantics
One advantage of spending decades in enterprise software is that you learn to distinguish between technology trends and architectural requirements.
Technology trends change.
Architectural requirements tend to endure.
Regardless of whether organizations were implementing ITSM, cloud platforms, workflow automation, analytics, or AI, the same challenge consistently emerged: how do you create a trusted understanding of the business across fragmented systems?
That challenge has not disappeared in the age of AI. In many ways, it has become more important.
The difference is that AI exposes inconsistencies much faster than traditional applications ever did.
If your business definitions are inconsistent, AI will amplify the problem.
If your data lacks context, AI will expose the gap.
If your knowledge is fragmented, AI will struggle to reason effectively across the enterprise.
That is why I believe every successful enterprise AI strategy must be built on a semantic foundation that connects, contextualizes, and governs enterprise knowledge.
Separating Signal from Noise
After more than three decades in enterprise software, I’ve learned that every major technology shift creates two things: genuine innovation and a tremendous amount of noise.
AI is no different.
Today, nearly every vendor claims to provide the context, intelligence, or data foundation required for enterprise AI. New categories are emerging almost monthly, and organizations are being inundated with competing messages about how to make their data AI-ready.
It reminds me of physical fitness.
Most people understand that exercise is important. Many start a program. Some stick with it for a few weeks. Very few build the discipline, consistency, and foundation required to achieve lasting results.
The same dynamic exists in enterprise AI.
Almost everyone agrees that AI needs context. Almost everyone agrees that trusted data matters. Yet far fewer organizations have invested in the semantic foundation necessary to deliver that context consistently across the enterprise.
The challenge is that context cannot be bolted on after the fact. It must be built into the way enterprise knowledge is connected, governed, and understood.
As AI agents become more autonomous and enterprise use cases become more consequential, that foundation will matter even more. Organizations will need AI systems that do more than retrieve information. They will need systems that understand business meaning, reason across complex relationships, and produce outcomes that can be trusted.
That is the work Stardog has been doing for more than two decades.
Stardog enables enterprise AI by providing the semantic layer and knowledge foundation that allow AI to understand, reason over, and trust enterprise data.
In the end, AI success is not simply a model problem. It is a knowledge problem. And the enterprises that solve the knowledge problem will define what AI becomes next.
