Data activation and Newton’s first law

My father has a PhD in mechanical social engineering and, among many things, designed cooling systems for nuclear reactors. From him, I inherited a restless curiosity about how the world works and a deep respect for the laws of science. Over the years, I’ve found that many of the principles governing physical objects translate perfectly to business.

Take Newton’s first law of motion. Objects at rest tend to stay at rest, and objects in motion stay in motion—at least in a vacuum. What’s true in physics is true for companies. Those that stall are destined to fall behind. But the businesses that remain in constant motion? They succeed.

In the agentic AI era, data is the force that overcomes business inertia. Period. Full stop. We often hear that data is our most essential asset, but that’s only a half-truth. To truly create value, critical information can’t just be pooled in a data lake somewhere or siloed in disconnected systems and walled gardens scattered throughout the organization.

Data has to move. It must be accessible, governed, timely, and most importantly, have the proper context for your business. Without all of that, there’s no use for an agent. The data might as well not exist.

When data moves, the business moves. That’s data activation.

THE MOUNTING DATA CHALLENGE

The amount of data in the typical enterprise is staggering and growing every day. In my most recent book, Digital Impact, I argue that we’ve fallen into the trap of thinking that all data is valuable, when in fact the vast amount of digitized information is “sand.” It’s unhelpful to our businesses, and the sheer volume has become a massive signal-noise problem.

The trick is being able to sift through all those grains of sand, identify the ones that really are flecks of data gold, and move them to where they can do the most good. It’s the same fundamental challenge that has been with us forever. It’s not new. But the stakes have risen. Solving it in the AI age makes finding a solution more essential than ever because high-quality data is the key to equipping autonomous and semi-autonomous agents with what they need to do their jobs.

The winners in business won’t be the ones with the most agents. Instead, they’ll be the ones with agents that are integrated best into critical systems, most deeply deployed in business processes, and powered by the most trusted data.

We’re all aware of sobering studies showing that in these early days of AI, most projects struggle to see meaningful ROI. Perhaps the biggest impediment businesses are facing is that they haven’t yet solved their foundational data problems.

Critical systems aren’t connected to the broader digital ecosystem, resulting in pockets of stranded data. That creates an inability to get the correct data to the right place—including to agents. Finally, there’s a lack of data governance, which undermines confidence in agents’ work. And as I’ve written before, change only happens at the speed of trust.

So, what should every business be thinking about when it comes to data activation? Consider asking three questions.

1. ARE YOUR SYSTEMS CONNECTED?

Trapped data fits neatly into Newton’s first law. It’s at rest, creating friction that prevents end-to-end workflows.

Critical applications have to connect to the broader technology ecosystem. You can’t activate what you can’t reach. Data orphaned in a legacy, on-premises ERP is dead weight. It’s the same with the proliferation of modern SaaS applications, which have added to the complexity and system fragmentation. These bottlenecks prevent end-to-end data workflows.

Creating a digital fabric through connectors and APIs makes all data available to every endpoint and puts it in motion for delivery to where it can have the greatest impact on the business.

2. DOES YOUR DATA HAVE CONTEXT?

Agents don’t just need data. They need to understand what it means—context. This is the “sand” problem, and it’s critical to solve it to get the most from agents.

Every agent starts “dumb.” Yes, it might be trained on generic data, but it knows nothing about your specific business. Feeding a raw database dump to an agent is insufficient. Activation isn’t just data movement deriving from connectivity. It’s the capability of sifting through the data sand—modeling, enriching, synchronizing—that makes it contextually relevant. It tilts the signal-to-noise ratio in your favor.

With agents, things can go bad quickly with inaccurate or insufficient data. That includes the well-documented issues of hallucinating and drifting. It’s also like the butterfly effect, where the downstream ramifications can get worse and worse. But when agents have access to complete, trusted, and timely data, they have the context they need to do their work accurately.

3. DO YOU HAVE GOVERNANCE?

Governance is the foundation of trust, and without it, businesses instinctively tap the brakes on AI initiatives. If you don’t have visibility into where your data is going, how secure it is, or how agents are using it, you won’t have confidence in the actions that agents are taking. That helps explain the monumental challenge we’re seeing of moving from pilots to production.

Governance doesn’t stop at visibility and security. Sovereignty is also a critical consideration. It’s about maintaining control over where data resides, how it’s stored and processed, who can access it, and how it’s contained within borders. As AI agents operate autonomously across systems and jurisdictions, organizations must be able to enforce data residency requirements, comply with evolving regional regulations, and clearly demonstrate accountability. Leaders need confidence that sensitive data isn’t being replicated, transferred, or exposed in ways that create risk.

Without this level of control, there is no trust, and without trust, AI initiatives stall before they ever reach scale. Effective governance means applying standards and guardrails to data movement to help you control agent behavior in real time. That allows you to deploy agents responsibly and with confidence.

ELIMINATE DATA FRICTION

Think about the millions of transactions that enterprises have to orchestrate every day, across thousands of systems, in real time, securely and reliably. That enterprise-level complexity isn’t going away.

Most companies are struggling to make sense of their data. They just want their companies to work better with the efficient flow of data through systems.

Activating trusted, contextualized data has always done that. Today, it also enables better decision-making by agents who work both on our behalf and with us.

Steve Lucas is chairman and CEO of Boomi.