Why AI fails without a strong data foundation and how to get it right

AI is transforming customer experience but most organisations are running in the wrong direction, chasing the newest tools while ignoring the foundation that actually drives results: their data.

A new survey from Harvard Business Review Analytic Services found a mere 7% of respondents claim their organisation's data is completely ready for AI adoption, and more than one-quarter (27%) report their data is not ready.

While organisations across industries have made sweeping investments in cleaning their data, they all recognise that even the most cutting-edge technology is only as powerful as the information that powers it. A fragmented, inconsistent data foundation doesn't just limit personalisation efforts; it makes every future AI investment costlier and less effective. 

The choice your organisation makes now about how to unify, govern, and activate your customer data comes down to a strategic decision that determines whether your AI ambitions deliver real ROI or remain on your roadmap wishlist.

The Power of Data

A topic that continues to come up in my discussions with CIOs and CTOs is this concept of "future-proofing." Specifically, future-proofing systems and architecture where the state of an organisation's data foundation will directly impact any planning discussions. 

In a recent conversation with a major UK food retailer, one that's been a digital-forward organisation for decades, the need for clean, actionable data was clear. This retailer sits on an extraordinary asset: over 50 years of rich customer data embedded deep within legacy systems on top of millions of records. The challenge is that a more modernised enterprise architecture has been built over top of that foundation and left their most valuable data siloed and underutilised.

To address this issue, think about the parameters when initiating change within any large organisation. It's important to ensure there's no friction introduced into downstream systems, and equally important, that the customer experience evolves without disrupting the trust members have already built with the brand. 

Plans should include architecting a future-proofed data foundation that doesn't just solve for today's operational burden, but unlocks the agentic automation capabilities that will reduce in-house overhead and strategically position organisations to leverage full personalisation capabilities and at scale. 

Unified at the Core

In another conversation with a global hair and skincare manufacturer managing nearly 20 brands across both direct-to-consumer and B2B channels, the challenge of delivering personalised customer experiences at scale quickly exposed the limitations of a fragmented data architecture. 

Each brand demanded its own unique customer journey tailored to its loyal audience, yet the underlying technology infrastructure was ill-equipped to support that level of individualisation without a unified foundation to build from. The tension between brand-level customisation and enterprise-wide consistency became a growing obstacle.

The solution lay in designing a technology strategy that could serve the organisation's largest, highest-volume brands first, then scale as needed. By standardising on a shared CRM, aligning marketing tools across the portfolio, and enforcing consistent data protocols across every brand, that created a foundation flexible enough to honour each brand's identity and strong enough to support the entire enterprise. 

This approach made clear that the path to meaningful personalisation at scale is not more technology - it is better, more unified data. For companies facing similar growing pains, and before investing in the next wave of AI-driven marketing tools, audit your data architecture and ensure that a unified foundation is in place to make those investments count.

The First Step

Ask these three questions before you create a technology strategy that focuses on building a strong data foundation for your AI programming:

1. What is the current state of your organisation's data? Before any technology strategy can take shape, you need an honest assessment of where your data actually stands today, including how it is collected, stored, and governed across your organisation. Identify where silos exist and which sources can be trusted. Without this clarity, any investment in AI tooling risks being built on unstable ground.
2. What tools and technologies have you already invested in? Taking stock of your existing technology landscape is essential before layering in new solutions, as many organisations already have underutilised platforms that could form the backbone of a stronger data strategy. Understanding what you have, what integrates well together, and where the gaps lie will prevent costly duplication and ensure that new investments complement rather than compete with what is already in place. 
3. Which teams have the capacity to drive and support? Even the most well-designed data architecture will fall short if the people responsible for maintaining it lack the time or skills to do so effectively. Map out which teams will own the data foundation, who will manage the tools day-to-day, and where training or additional resources may be needed to close capability gaps. A technology strategy is only as strong as the people empowered to execute it.

Answering these questions honestly and thoroughly is not only a useful exercise, but will help put your organisation on a path to succeed. Companies that take the time to get their data foundation right will be able to unlock the full potential of their AI investments and position themselves to adapt and lead as the technology continues to evolve.

An Important Investment

Most organisations are still making technology investments based on a 5-10 year time horizon. Or, even worse, anchoring them to processes that were designed twenty years ago when they started their business. It's the wrong approach, for many reasons.

Rather than optimising incrementally for today, organisations need to ask a more pressing question: what does the business look like through the lens of an agentic enterprise? From that vantage point, the priority becomes clear and it's time to define and build a data foundation that can support autonomous, intelligent decisioning at scale.

Because this is the shift: personalisation and AI are no longer programmes to deliver, but capabilities that define how the business runs. Organisations that invest with that mindset will shape what comes next.