NTT DATA Highlights Governance and Data Quality as Keys to Scaling Enterprise AI

Artificial intelligence adoption is expanding across businesses, but many organisations are struggling to turn successful experiments into systems that deliver measurable results at scale. A recent report from NTT DATA highlights a persistent gap between AI pilot projects and company-wide deployment, pointing to data quality, governance, security and business process redesign as essential elements of successful implementation.

The report found that 59% of surveyed respondents said their AI initiatives remained limited to small pilot projects or were deployed within individual business functions without being integrated across the organisation. The finding illustrates how companies can make progress with isolated AI applications while still struggling to establish a consistent approach across their operations.

For enterprises seeking greater returns from AI investments, the challenge is no longer simply choosing a model or introducing a new tool. It is building the organisational and technical foundations needed to make AI work reliably within everyday business processes.

Why AI Pilot Projects Struggle to Scale

An AI system that performs well in a controlled test environment may face very different demands when introduced to thousands of employees, connected to existing applications and used in business-critical decisions.

Jesus Ordonez Garcia, CEO of consulting for Benelux and France at NTT DATA, has highlighted the difficulty of moving from a small, successful experiment to a system that can operate across an entire organisation. Data integration, security and governance become increasingly important as the number of users and connected processes grows.

Businesses must also determine how AI applications will interact with their existing technology infrastructure. Separate projects may use different datasets, platforms and operating rules, making it difficult to share information or coordinate activities effectively.

Without a common strategy, organisations risk building multiple AI tools that solve individual problems but fail to contribute to a broader business objective.

Data Quality and Security Remain Critical

Reliable data is a fundamental requirement for effective AI deployment. Systems trained or operated using incomplete, inconsistent or inaccurate information can produce poor results, regardless of how advanced the underlying technology may be.

Jeroen van Hamersveld, managing director of NTT DATA Netherlands, has emphasised the need for organisations to understand the quality of their data before expanding AI usage.

Companies must also decide how information will be stored, processed and protected. Some applications may require AI to operate within private data environments, while others may permit broader access to external services. These decisions affect privacy, security, compliance and an organisation's control over its information.

Establishing clear rules early can help businesses avoid costly changes later, particularly when AI applications begin handling sensitive information or supporting important operational decisions.

Governance Must Include Employees and Business Processes

Technology alone cannot ensure successful AI adoption. Employees need training, clear responsibilities and practical guidance on how to use AI tools appropriately.

Van Hamersveld has argued that workers should have opportunities to experiment with AI, while organisational governance ensures that useful experiments translate into business value. This approach seeks to balance innovation with oversight rather than restricting experimentation altogether.

Business process redesign is equally important. Applying AI to an inefficient workflow may simply accelerate an existing problem instead of improving the outcome. Organisations therefore need to examine how work is performed before deciding where automation or AI assistance can deliver meaningful benefits.

As AI systems become more autonomous and influence business decisions, companies also need defined accountability, monitoring and controls.

Coordinating AI Agents Across the Enterprise

Another challenge is ensuring that individual AI agents work together rather than operating as disconnected tools. Coordinated systems can support broader workflows, provided their responsibilities, access permissions and interactions are clearly managed.

NTT DATA positions itself as an orchestration partner, helping businesses connect AI initiatives with their wider technology infrastructure and strategic objectives. The approach reflects a broader shift towards treating AI as an organisation-wide transformation rather than a collection of separate software deployments.

Ultimately, scaling enterprise AI requires more than expanding access to new tools. Companies must align reliable data, secure infrastructure, employee readiness and accountable governance with clearly defined business goals. Those foundations will help determine whether AI experiments become sustainable capabilities that improve everyday operations.