This is article four in a six-part series exploring new research findings from HCLTech and Raconteur. You can read the rest of the article series in ‘The Blueprint for AI Leadership’ here or download the full report here.
As organisations move from building AI readiness to scaling it in practice, attention is shifting to a more fundamental constraint: the state of their data.
Once AI initiatives expand beyond pilots, gaps in data quality, accessibility and governance become increasingly difficult to ignore. AI is only as powerful as the data it runs on, yet a growing divide is emerging between organisations that have built strong data foundations and those that have not. Leaders are eight times more likely than Followers to express confidence in their data foundations when supporting GenAI initiatives (74% vs. 9%), a gap that is increasingly shaping which organisations can scale AI and which remain stuck in pilot mode.
This disparity is also reflected in performance outcomes, suggesting that data maturity is becoming a key marker of which organisations can turn AI ambition into business value.
Who are AI Leaders & Followers?
AI Leaders, who represent 18% of the sample, are maximising ROI not simply by using AI to reduce costs, but by applying it to drive growth, innovation and improved customer experience.
AI Followers, by contrast, represent 60% of the sample. While they are realising returns from AI investments across business functions, they significantly trail Leaders in the higher-value use cases that are reshaping operating models and competitive advantage.
From AI ambition to data reality
The conversation around AI has shifted significantly in recent years. “A couple of years ago, everyone was focused on the models themselves. Today, the differentiator has become abundantly clear – it’s the data,” says Nagaraj Sastry, senior vice president and global head, data and AI, digital business services at HCLTech. “The technology has become commoditised in many ways. But what hasn’t been commoditised and what can’t be easily replicated is your proprietary data, your institutional knowledge and, crucially, your ability to leverage it.”
For many organisations, this is where the challenge becomes most visible. “Every organisation thinks AI is a magic tool that’s going to solve all their problems, but it’s really a magnifying glass that exposes problems in their underlying data,” says Marvin Ward Jr., global head of data management and operations at Bloomberg.
Despite continued investment, many AI programs fail to move beyond proof of concept. The research points to a common underlying cause: fragmented data ecosystems.
These environments are often the result of years of decentralised decision-making, leaving organisations with siloed datasets, inconsistent definitions and limited interoperability. The barriers, however, are not purely technical. “Data has historically been treated as a departmental asset,” says Sastry. “Each group has built systems optimised for their specific needs, but they’ve created a fragmented landscape that’s genuinely difficult to unify.”
Trust is another critical constraint. organisations frequently lack confidence in the accuracy, completeness and timeliness of shared data, while internal dynamics can actively discourage collaboration. At the same time, legacy infrastructure compounds the issue. Many systems were not designed to interoperate, creating integration challenges that are complex and costly to resolve. The result is a data environment that remains ill-equipped to support enterprise-scale AI.
AI has drastically lowered the barrier for users to query an organisation’s data
As Ward Jr. notes, “AI has drastically lowered the barrier for users to query an organisation’s data. This makes data interoperability a must-have and this is where many organisations struggle.”
Even where organisations successfully generate insights from AI, the ability to translate them into action remains uneven. Many say their confidence is highest when AI is used to support operational performance recommendations (38%), predictive forecasting (38%) and productivity optimisation insights (31%). However, this confidence drops as organisations move closer to automation: just 22% express high confidence in automated decision-making and only 17% in agent-initiated actions.
What emerges is that many organisations are comfortable using AI to inform decisions, but far fewer trust it or have the data foundations required to operationalise those decisions at scale.
What distinguishes AI Leaders
Against this backdrop, AI Leaders stand apart not because they have access to better models, but because they have invested earlier and more systematically in their data foundations. “These organisations did not wait for AI to become a priority before addressing data challenges. Instead, they invested in infrastructure, governance and quality over time,” says Sastry.
Leaders treat data as a managed product, with clear ownership, defined standards and lifecycle management. 51% of AI Leaders have clear ownership and accountability for data-driven AI decisions, while only 27% of AI Followers do. Data is curated for reuse rather than generated for one-off applications. Leading organisations invest in data observability, enabling continuous monitoring of quality and real-time detection of issues, ensuring that confidence in AI systems is grounded in confidence in the data that feeds them.
Architecture plays a critical role in the success of Leaders who prioritise interoperability and scalability over isolated solutions, whether through modern platforms or distributed models such as data mesh. Equally important is discoverability. Investments in metadata and data cataloging make it easier for teams to find, understand and use data across the organisation. As Ward Jr. explains, “Firms must make explicit investments in their metadata to tie together relevant data across domains.”
[AI Leaders have] invested in infrastructure, governance and quality over time
There is also a human dimension that organisations must get right. AI Leaders invest in data literacy beyond technical teams, establish clear stewardship models and create forums for collaboration between data producers and consumers, ensuring that data is not only accessible but actively used across the organisation.
In turn, this is reflected in how data is shared. AI Leaders are far more likely to operate in environments where data can move across teams and functions: nearly half (46%) report having a fully established, responsible and trust-based data-sharing culture, compared with a fifth (21%) of Followers.
This does not mean granting unrestricted access but rather developing a culture in which data is well understood and therefore easier to use consistently across domains. As Ward Jr. notes, “Data without borders is actually more about data consistency than governance and security.”
This consistency is a critical piece of any organisation’s ability to scale AI. The most valuable use cases require combining data from multiple functions, including customer intelligence, operational optimisation and risk modeling. Without the ability to connect these domains, AI remains limited to narrow, localised applications.
Crucially, culture underpins all of this. “You can have the most sophisticated technology stack in the world, but if your organisation operates with a mindset of data hoarding, you’ll never realise the full potential of enterprise AI,” says Sastry. Creating this culture requires deliberate leadership action. Incentives must be aligned with enterprise-wide outcomes, encouraging teams to share and contribute data rather than retain control.
It also requires new ways of working, with cross-functional collaboration embedded into organisational processes and supported by data communities and centres of excellence. Trust must be built at multiple levels – technical, organisational and cultural – so that teams have confidence in data quality, clarity on how it will be used and assurance that their contributions will be recognised.
Governance as an enabler, not a constraint
As organisations seek to scale AI, governance becomes increasingly important, particularly in the context of GenAI, where risks are more visible and more complex. “Governance isn’t the enemy of innovation; it’s what makes sustainable innovation possible,” says Sastry.
While safe and compliant data use is relatively well established among Leaders (64%), the capabilities that make AI trustworthy at scale remain less mature. Only a third of Leaders say that the explainability of AI outputs (35%) and training and literacy on Responsible AI practices (32%) are well established. Both measurements drop to less than one in six for Followers.
Effective governance, therefore, extends beyond basic compliance to ensuring that data is accurate and well-structured while embedding transparency, accountability and safeguards into AI systems.
For organisations seeking to close the gap between AI ambition and data readiness, the path forward is rarely linear. “The first question to ask is: what’s actually blocking you today?” says Sastry. For some, the issue is visibility: understanding what data exists and where it resides. For others, it is legacy infrastructure that cannot support modern AI workloads. In many cases, governance gaps either introduce risk or restrict access.
The research highlights a clear disconnect. While strategic integration is widely recognised as critical to AI success, just one-third believe they are investing sufficiently in the foundations required to enable it. Nearly half (49%) cite data-related challenges as a barrier, while 47% point to technology limitations. In other words, many organisations understand what matters but are underinvesting in the systems required to deliver it. As Ward Jr. explains, infrastructure, governance and culture must be addressed together: “think of this like a three-legged stool.”
Data readiness, like change management, is not a one-off initiative but an ongoing organisational capability. “Many organisations are still treating AI initiatives like traditional IT projects,” says Sastry. “But AI doesn’t work that way. It requires continuous refinement.”
Each improvement reinforces the next. Better governance enables faster decisions, improved architecture increases accessibility and greater accessibility drives usage – generating feedback that further improves quality. In this way, organisations that succeed will treat data not as a prerequisite to AI, but as the foundation on which it continuously evolves.
For business leaders, the priority areas of focus should be strengthening governance, improving accessibility and building a mature culture in which data is shared, trusted and strategically managed. Without trusted, accessible and well-governed data, even the most advanced AI initiatives will struggle to move beyond experimentation.
That said, strong data foundations alone are not enough. As organisations attempt to operationalise AI at scale, they must confront the challenge of whether their underlying architecture can support it.
This is article four in a six-part series exploring new research findings from HCLTech and Raconteur. You can read the rest of the article series in ‘The Blueprint for AI Leadership’ here or download the full report here.
This is article four in a six-part series exploring new research findings from HCLTech and Raconteur. You can read the rest of the article series in 'The Blueprint for AI Leadership' here or download the full report here.
As organisations move from building AI readiness to scaling it in practice, attention is shifting to a more fundamental constraint: the state of their data.
Once AI initiatives expand beyond pilots, gaps in data quality, accessibility and governance become increasingly difficult to ignore. AI is only as powerful as the data it runs on, yet a growing divide is emerging between organisations that have built strong data foundations and those that have not. Leaders are eight times more likely than Followers to express confidence in their data foundations when supporting GenAI initiatives (74% vs. 9%), a gap that is increasingly shaping which organisations can scale AI and which remain stuck in pilot mode.
This disparity is also reflected in performance outcomes, suggesting that data maturity is becoming a key marker of which organisations can turn AI ambition into business value.
