AI without the data fundamentals: Turbocharge your way to ruin or success?

Leadership teams are approving ambitious AI budgets, fascinated by the possibilities of generative AI, automation and intelligent tooling and there is no doubt that this is happening in boardrooms everywhere. The demonstrations are impressive, the interfaces are slick and the potential productivity gains can be compelling.

But there is a less exciting question that needs to be asked before organisations start investing heavily in the latest technology: Is the data underneath it actually fit for purpose?

For many organisations, the uncomfortable answer is: not yet, because there is a danger that businesses move into 2027 on the AI hype train, investing heavily in new platforms while remaining burdened by exactly the same data problems they have struggled with for years. AI will not make those problems disappear. In some cases, it could make them considerably more damaging.

The shiny technology problem

Part of the challenge is understandable because AI is all too easy to get excited about. When you add a polished demonstration showing a platform answering questions in seconds, automating processes, producing analysis, identifying trends or transforming the way employees interact with information then the client is ready to buy.

However, a data strategy rarely produces the same reaction. Data governance, architecture, ownership, quality, security and establishing a reliable single source of truth is unlikely to generate the same excitement in a boardroom because this is information which is foundational rather than flashy.

Data determines whether the exciting technology sitting on top actually works and you can invest in the most sophisticated AI tools available, but if they are accessing fragmented, duplicated, outdated or inaccurate information, you have simply given bad data a much more powerful delivery mechanism.

Rubbish in. Rubbish out. Just much faster.

Recruitment has long had a simple expression when talking about databases: rubbish in, rubbish out and this principle applies perfectly to AI. Giving unstable or inaccurate data to an advanced AI platform does nothing to resolve the underlying problem. Instead, there is a risk of turbocharging it.

Suddenly, incorrect numbers, inconsistent records and unreliable metrics can be processed, summarised and distributed at extraordinary speed. Worse still, they can be packaged within an impressive conversational interface that makes the information appear authoritative.

The technology may be new but the underlying information problem is not. The real difference is scale. Where poor data might previously have created an inaccurate report or required someone to manually reconcile two conflicting spreadsheets, AI has the potential to propagate that information across workflows, analysis and decision-making far more quickly.

That turns data quality from an IT inconvenience into a genuine business risk.

AI needs something reliable to stand on

Artificial intelligence is an enormously powerful capability, but it is not a substitute for getting the fundamentals right. Before asking what AI can do for the organisation, businesses need to understand the information they expect those tools to consume.

Where does the data come from? Who owns it? How accurate is it? Is the same information held differently across multiple systems? How securely can it be accessed? And, crucially, which source should the organisation actually trust?

A strong data strategy provides those foundations and that means establishing clear ownership and governance, improving accuracy and consistency, removing unnecessary duplication, understanding how information moves between systems and creating trusted sources from which AI tools can operate. None of this is particularly glamorous, but neither are foundations when you are building a house.

From AI experimentation to business value

There is also an important distinction between experimenting with AI and embedding it into the way an organisation operates because the former can happen quickly, but sustainable transformation takes more work.

As AI moves beyond isolated pilots and becomes connected to operational systems, customer information, financial data and management reporting, the quality of the underlying data becomes increasingly important and therefore, the objective should not simply be to introduce more AI, but to create an environment in which AI can improve the quality and speed of decision-making without compromising the reliability of the information on which those decisions are based.

Get that right and the opportunity is enormous because AI can help organisations interrogate complex information faster, automate repetitive processes, identify patterns that would previously have been difficult to uncover and give employees easier access to the information they need.

But those benefits depend on trust and if users continually have to question whether the information produced by an AI platform is accurate, much of the promised productivity gain disappears.

The conversation organisations need before 2027

As technology strategies and budgets for 2027 are developed, perhaps the first conversation should not be about which AI platform to buy but instead should be about whether the organisation is ready for it. That definitely does not mean businesses should delay innovation until every historical data issue has been resolved,  because few organisations will ever reach that point.

It does mean that AI investment and data strategy need to progress together. Bypassing the data foundations risks creating impressive technology that simply repeats historical mistakes back to the organisation. The errors will arrive faster, look more sophisticated and potentially be delivered with far greater confidence.

The goal is to build on reliable foundations, however, and the picture changes completely. With trusted, accessible and well-governed data beneath it, AI has the potential to transform how organisations operate in 2027 and beyond.

Without those foundations, you may still turbocharge the business.

Just not necessarily in the direction you intended.

Get in touch

If you are in the Digital, Data and Technology space, either client and candidate, please email me on david.flynn@merakitalent.com or meet me on LinkedIn.