Organisations that slow down can get ahead in the AI race amid greater regulation

AI is often framed as a race. The faster an organisation adopts the technology, the further ahead it will be.

But as key provisions of the EU AI Act come into force, businesses are being reminded that successful AI adoption depends on more than speed. Governance, knowledge management and data quality can no longer be treated as issues to address later. The organisations that gain the most value from AI may not be those moving first, but those taking the time to build the right foundations before scaling their ambitions.

For many leaders, the challenge is no longer deciding whether to use AI. It is understanding how to adopt it in a way that is responsible, sustainable and capable of delivering long-term value.

AI cannot compensate for missing knowledge

Much of the excitement surrounding AI centres on what the technology can do. Yet far less attention is paid to the quality of the information it relies on.

For many organisations, crucial operational knowledge remains fragmented across teams, stored in outdated systems or held in the minds of experienced employees. AI can help organisations access and apply knowledge more efficiently, but it cannot create knowledge that has never been captured in the first place.

This challenge helps explain why many AI initiatives struggle to deliver on their initial promise. Research from RAND found that more than 80% of AI projects fail, often due to issues such as poor data quality, unclear objectives and organisational readiness. The technology itself is rarely the problem. More often, organisations are attempting to build AI-enabled services on foundations that were never designed to support them.

Before investing in increasingly sophisticated AI capabilities, businesses should take a closer look at how knowledge is created, shared and maintained across the organisation. Without that groundwork, even the most advanced AI tools will struggle to deliver meaningful value.

The AI Act raises the stakes

The rush to adopt AI has largely been driven by a desire to improve productivity. The EU AI Act adds another consideration: accountability.

As key provisions come into force, organisations are under growing pressure to understand how AI systems are used, what data they rely on and how decisions can be explained when customers, regulators or employees ask questions. Transparency requirements mean certain AI-generated content must be clearly identified, while organisations using higher-risk systems face stricter obligations around governance and oversight.

Many businesses, however, are still treating governance as an afterthought. They have started experimenting with AI agents and automation tools before fully understanding the processes, data flows and knowledge repositories those systems rely on. That creates both operational and compliance risks.

AI governance cannot be bolted on later. Organisations need visibility into how information is collected, managed and used from the outset. Increasingly, the ability to demonstrate that control is becoming just as important as the AI capabilities themselves.

Start small and make the value visible

AI's potential is vast, but that doesn't mean every organisation should begin with ambitious transformation programmes. In many cases, the most successful AI initiatives are those that start with clearly defined problems and measurable outcomes.

Rather than attempting to automate entire business processes, organisations should focus on targeted applications where value can be demonstrated quickly. Tasks such as summarising information, improving written content, translating communications or helping employees retrieve knowledge can deliver immediate benefits while allowing teams to build confidence in the technology.

This approach has another advantage. Smaller, well-scoped projects make it easier to understand how AI interacts with data, workflows and governance requirements before wider deployment. They provide an opportunity to identify risks, refine processes and establish best practice in a controlled way.

The organisations seeing the greatest success with AI are not necessarily those making the largest investments. More often, they are the ones taking a deliberate approach, learning from early use cases and scaling from a solid foundation. Long-term success with AI is built through a series of practical improvements, not a single transformational leap.

Human insight remains a competitive advantage

AI can accelerate tasks, surface information and automate routine work, but it cannot replace the knowledge, judgement and context that people bring to an organisation. In fact, as AI becomes more embedded in everyday business operations, human expertise becomes even more important.

The success of any AI initiative ultimately depends on the quality of the information behind it and the decisions made about how it is used. Experienced employees provide the context that helps organisations interpret results, challenge assumptions and identify when something is missing or simply doesn't make sense. Those are capabilities that remain difficult to automate.

This is why organisations should view AI as a tool to augment people, rather than a substitute for them. The greatest value often comes from combining the speed and efficiency of AI with human insight, oversight and expertise.

As organisations continue to navigate the evolving regulatory landscape and growing pressure to adopt AI, the temptation will be to move faster. But long-term success is unlikely to belong to those who deploy the most AI. It will belong to those who combine the technology with strong governance, well-managed knowledge and the human insight needed to use it effectively.