Release the Handbrake: Unlocking AI Through Solid Data Foundations

In financial services — and across many industries — technological innovation often arrives with great promise. We’ve seen waves of enthusiasm for new tools and platforms that were billed as game-changers, only for interest to wane after limited real-world impact.  

While a few early adopters might extract value, widespread uptake is typically hampered by organisational caution or a lack of readiness, and the innovation quietly fades into history. 

Over the past few years, however, one trend has endured: artificial intelligence (AI). Unlike past “false dawns,” AI is not just another fleeting headline — it is transformative, it is evolving, and it is here to stay. 

AI, like many of its technological predecessors, promises increased efficiency, enhanced optimisation, and improved service delivery. But these outcomes depend on something far less glamorous — a solid data foundation, including:  

• • A well-defined approach to data governance. 
  • An understanding of end-to-end processing, including the acquisition, utilisation and retention of data. 
  • An effective blend of preventative and detective operational controls that align with organisational risk appetites. 
  • An understanding of the difference between data integrity (is data complete and does it meet basic format requirements) and operational data quality (is the data fit-for-purpose). 
  • Appropriate role-based enterprise data literacy to enhance the effectiveness of how data is utilised. 

For AI to be more than a theoretical advantage, it must be built on data that is accurate, complete, well-governed, and trusted — as the adage saying goes; “rubbish in, rubbish out”. 

The Cost of Neglecting Data Quality 

Gartner predicts that 80% of AI initiatives will fail by 2027 due to poor data quality. This figure should give every leader pause. It highlights a paradox: while AI is the focus of significant investment and high expectations, success often hinges on disciplines that remain underfunded and undervalued. 

Rather than viewing data governance and quality as barriers or afterthoughts, organisations must recognise them as enablers of innovation. These defensive capabilities are not optional; they are prerequisites for reliable, scalable, and ethical AI. 

The Hidden Tension: Offense vs. Defense in Data Strategy 

There is a persistent tension in how organisations approach their data strategy and the associated investment decisions. This tension often arises from viewing each element of the strategy in isolation, rather than as interconnected components that enhance and reinforce one another. In reality, success relies not on the individual value of each element, but on the collective impact they deliver as a cohesive whole. 

Generally speaking, there are two core data utilisations:  

The offensive side — focusing on innovation, analytics, and AI — which receives attention and funding because it is exciting, future-focused, has quantifiable return of investment and for those reasons, is easy to sell.  

In contrast, the defensive side — data quality, governance, and enterprise approach to data literacy and risk awareness — is often deprioritised, dismissed as operational overhead or viewed as a compliance obligation. 

However, this way of thinking can at times be misleading. The offensive and defensive aspects of data strategy are not opposing forces — they are mutually reinforcing. As the old sporting adage goes: “Offense wins games, but defense wins championships.”  Whilst we may not be running onto the MCG on the last Saturday in September, the same principle applies to data and AI.  

You cannot fully capitalise on cutting-edge technologies and realise the anticipated benefits that the offensive side of data can deliver without addressing defensive side of data, or an understanding of data-related vulnerabilities and how to mitigate them. 

Changing the Narrative: From Compliance to Value Creation

To shift organisational thinking, we need to reframe data quality and strategy not as a cost but as a strategic asset for the entire organisation. Yes, it can be difficult to quantify the ROI of better data governance or cleaner datasets. But the benefits are clear:

• Reduced remediation and rework
• Lower operational and reputational risk
• Stronger client trust and satisfaction
• Increased confidence in decision-making
• Higher likelihood of success in AI and analytics initiatives
• Improved operational data utilisation in the broader context of regulatory obligations

These are not soft benefits — they are essential to long-term business performance.

A Cyclical Journey, Not a Linear One

The relationship between defensive and offensive data investment is not linear — it’s cyclical. Strong data governance enables effective innovation (do I hear you say ‘AI initiatives’?!), which in turn reveals further opportunities to improve and optimise data. Organisations that recognise this cyclical dynamic can create a sustainable feedback loop between foundational capability and strategic ambition.

Moreover, investment in AI presents an opportunity to address long-standing gaps in data management. As excitement and funding grow for offensive capabilities, leaders should use this momentum to uplift their data foundations — ensuring AI initiatives are built on solid ground.

Time to Release the Handbrake

This is a pivotal moment. AI will redefine the way financial services operate, compete, and create value. But its success is not guaranteed. Without investment in solid data foundations, even the most advanced AI tools will fall short.

The handbrake will look different within each organisation based on their industry, maturity and risk appetites, but there are several items that can be consistently considered when deciding to release it:

• Do we have confidence in the completeness and quality of data that will be utilised within our AI strategy?
• Is there a holistic approach to organisational data strategy that recognises the attainment of anticipated benefits, in part, is impacted by controllable items?
• Are there effective operational controls across the end-to-end data supply chain?

To unlock AI’s full potential, organisations must release the handbrake. That means embracing and committing to the less glamorous, but critical, components of the data lifecycle. Because in data, as in sport, you don’t win championships without great defence.

Release the Handbrake: Unlocking AI through solid data foundations is part of The Quarterly – Q4 FY25

Key Contributor:

This article was also strengthened by a wider group of Novigi specialists, whose withering years of toil and rich experience added depth and clarity to the perspectives shared.

For more information about anything you’ve read here, or if you have a more general inquiry, please contact us.