Why AI Value Breaks at the Boundaries

The overlooked role of data movement in scaling enterprise AI

Artificial intelligence is rapidly becoming a core component of digital business strategies. Organizations across industries are investing in AI to improve operational efficiency, enhance customer experiences, support decision-making, and unlock new opportunities for innovation. As adoption accelerates, discussions naturally focus on models, platforms, governance frameworks, and regulatory requirements.

These are all important considerations. However, there is another factor that increasingly influences whether AI initiatives can deliver value at scale: the movement of data.

Most enterprise AI systems depend on data travelling between users, applications, cloud platforms, operational systems, partners, and locations. As digital operations become more distributed, these data flows become more complex, creating dependencies that are often difficult to see and even harder to manage. The conversation around AI frequently focuses on where data is stored and how models are governed. Increasingly, organizations must also consider how data moves.

After all, AI platforms, cloud services, cybersecurity controls, and digital applications all depend on the same foundation: data moving predictably, securely, and resiliently across interconnected infrastructure.

Why it matters now

Many organizations have successfully demonstrated the potential of AI through pilot projects and targeted use cases. The next challenge is scaling those successes across the enterprise.

In practice, the obstacle is often not the model or algorithm itself. It is ensuring that the right data can move to the right place, under the right controls, at the right time. AI systems rely on information that frequently crosses organizational, cloud, geographical, and jurisdictional boundaries. As a result, the effectiveness of AI becomes increasingly dependent on the infrastructure that enables those connections.

This raises an important question for business leaders: Where do your organization's most important data streams cross boundaries that you do not fully control?

The overlooked layer of digital operations

Cloud services, AI platforms, cybersecurity controls, and digital applications all rely on one common foundation: data moving predictably, securely, and resiliently across interconnected infrastructure.

Yet networks often receive less strategic attention than the technologies they support.

For many organizations, connectivity is viewed primarily as supporting infrastructure—something that simply enables users to access applications and services. As digital operations become increasingly distributed, that perspective is beginning to change.

The network is no longer just a utility. It is becoming a strategic dependency that influences how effectively organizations can operate, innovate, and scale emerging technologies such as AI.

Food for thought: Do you treat your network as a strategic dependency—or is it still viewed as background plumbing?

Why AI value breaks at the boundaries

AI creates value by connecting information from multiple sources.

As organizations scale AI initiatives, data increasingly moves between business units, cloud providers, partner ecosystems, operational technology environments, edge locations, and geographically distributed teams.

Every boundary introduces additional complexity. Governance requirements may differ. Visibility may decrease. Dependencies may become less obvious. Operational control may become more difficult to maintain.

These challenges are not unique to AI. However, AI amplifies them because its effectiveness depends on the ability to access, process, and act on data across organizational, cloud, and geographical boundaries.

The result is that, for many organizations, challenges in scaling AI emerge at the boundaries between systems, organizations, and infrastructure rather than within the models themselves.

The assumption that can create risk

Organizations often assume that if applications are hosted within a particular region, the associated data remains within that region. The reality can be more complex.

While private networking architectures can provide greater control over traffic paths, a significant proportion of enterprise traffic today traverses internet-based infrastructure, cloud interconnections, and third-party networks. Consider an organization with offices across Europe that hosts applications in a European cloud environment. Many stakeholders would reasonably expect traffic to remain within Europe.

However, internet routing does not necessarily follow geopolitical borders. It follows network topology, peering arrangements, and routing policies. As a result, traffic may traverse networks, providers, or jurisdictions that were never considered during architecture planning.

The destination may be exactly where intended. The path may not be.

For organizations seeking to scale AI while maintaining governance, resilience, and operational control, understanding these paths becomes increasingly important.

Food for thought: Do you know the typical path your most sensitive traffic takes—or only its intended destination?

Making dependencies visible

The objective is not to eliminate every dependency.

Modern digital operations depend on cloud providers, network operators, trust services, partners, and technology platforms. These dependencies are often necessary to achieve scale, flexibility, and innovation.

The challenge is ensuring that they remain visible, intentional, and manageable.

Many organizations have detailed visibility into applications and cloud platforms while having comparatively less visibility into the networks and paths that connect them. Areas that increasingly deserve attention include:

• Dependencies on external network providers
• Concentration risk associated with individual carriers or platforms
• Unintended cross-border routing
• Trust services such as DNS and certificate authorities
• Routing disruptions and unintended traffic paths

The issue is not whether these dependencies exist. The issue is whether organizations understand them well enough to manage them effectively.

From connectivity to operational control

Historically, enterprise networks were designed primarily to provide connectivity. Increasingly, organizations are looking to their networks not only for connectivity, but also for visibility, resilience, and operational control across distributed operations.

As AI becomes embedded within business processes, organizations need greater confidence in how critical data moves across their digital infrastructure. This requires visibility into traffic flows, routing behaviour, operational dependencies, and resilience measures.

Capabilities such as route monitoring, encryption in transit, multi-provider resilience strategies, local processing, and enhanced observability can help organizations gain greater insight into the infrastructure supporting AI-driven operations.

The goal is not simply to connect users and applications. It is to provide greater control over the movement of critical business data.

Food for thought: What would change if your network was designed as a control plane for digital operations?

Questions leaders should ask

As AI initiatives move from experimentation to operational deployment, leaders should consider several important questions:

• Which data flows are critical to our AI strategy?
• How much data needs to move between systems, locations, and partners?
• How latency-sensitive is that data?
• Which dependencies are acceptable—and which would be difficult to explain after an incident?
• Where do our most important data streams cross organizational, cloud, jurisdictional or geographical boundaries?
• Which business services would fail first if a provider, route, or jurisdiction became unavailable?
• How quickly would we detect disruptions affecting critical data paths?
• Where do we need greater visibility, resilience, or operational control?

For many organizations, answering these questions is the first step toward building a stronger operational foundation for AI at scale.

From awareness to managed resilience

The conversation around AI often focuses on algorithms, computing power, and governance frameworks. These topics remain essential.

However, as organizations seek to operationalize AI across increasingly complex digital operations, another question deserves equal attention: Can we see, understand, and influence how the data that powers AI moves across our digital infrastructure?

The goal is not to remove every dependency. It is to make dependencies visible, intentional, and resilient. Because the challenge is no longer whether AI can scale. The challenge is whether organizations retain sufficient visibility and control over the infrastructure that enables it.

Continue the conversation

As organizations scale AI initiatives, understanding how critical data moves across cloud platforms, locations, partners, and digital operations is becoming increasingly important.

If you would like to discuss how visibility, resilience, and data movement may impact your AI strategy, our specialists are available to help.