Briefing Frontiers 2025: AI can't deliver unless firms deal with data disconnects

The year 2025 is a pivotal one for proving the value of AI investments. After freely spending on AI in 2024, it’s clear there is pressure to demonstrate tangible business benefits. This year in particular, business leaders and organisations will be held accountable for the returns their investment in AI delivers.

A strong data strategy must underpin AI for organisations to successfully deploy and achieve a healthy ROI from their AI initiatives. However, not only are organisations (nearly 66%) investing in AI without a proper data strategy in place, there is also confusion and a lack of understanding about what constitutes a data strategy.

Merely having the role of chief data officer — as claimed by 39% of the respondents — doesn’t necessarily mean the organisation has, or will have, an effective data strategy. One can draw parallels with GDPR officer roles, which have proved to be tick-box assignments in many firms

From the report, it is apparent many organisations claim to have a data strategy, but don’t know what a data lake is, or even where data is stored. This calls into question their grasp of the subject. A data lake is essentially the ‘end’ of a data strategy — the common platform where an organisation puts all its data from different sources together to enable a centralised view of the information. We can see that over 80% of respondents claim they either have a documented and agreed-on data strategy for their firm, or will have one within 12 months. In comparison, only 34% either have a data lake today or are building one, and 47% have no plans, or are unsure if they will have one. These numbers don’t stack up.

Likewise, merely having the role of chief data officer — as claimed by 39% of the respondents — doesn’t necessarily mean the organisation has, or will have, an effective data strategy. One can draw parallels with GDPR officer roles, which have proved to be tick-box assignments in many firms to help comply with the requirements.

All these highly noticeable disconnects indicate wider issues around data management maturity. In the absence of robust data assimilation, curation and ongoing management, organisations will struggle to prove AI’s value to the business. These solutions need access to clean and organised data. A robust data strategy is an imperative and foundational requirement, without which AI initiatives are at risk of falling short of expectations in 2025.