Three trends that will drive more AI value in 2026

Behind all these complex labels lies a fairly simple change: organisations are moving away from focusing solely on the model and are paying more attention to the system around it – data, workflows, governance, people and decision‑making.

We will see this international trend reflected in the Netherlands as well. The question for executive teams is no longer whether they should “do something with AI”, but how they can embed AI so deeply that the organisation truly misses it when it stops working.

1. From personal productivity to critical value streams

Over the past two years, many Dutch organisations have followed a similar path. Microsoft Copilot licences have been rolled out, internal chatbots are increasingly common, and training programmes have been launched to help individual employees get started. The initial gains from this individual use are, however, limited: employees save roughly 15 to 30 minutes of time per day on repetitive tasks.

Personal productivity turns out not to be a benchmark for corporate AI success. It is difficult to measure and appears underwhelming at first glance. The real value – and where organisations will (and must) focus this year – lies in structurally embedding AI in the processes that generate business value. So not asking, “How many staff hours do I save each day?” but asking questions such as:

“What percentage of our costs and revenue is touched by AI each day?”

“Which decisions are we demonstrably making better or faster?”

“Which legacy ways of working have we actually replaced?”

When organisations steer towards these types of questions, the AI roadmap immediately becomes more strategically valuable. At the moment, it often resembles a laundry list, with dozens of isolated use cases for each team, each requiring its own technology and governance. Stacking up separate pilots does not create control over AI; organisations need a scalable and integrated roadmap with consistent criteria for data, risk, technology and adoption.

The outlines of this shift are already visible in large Dutch organisations. A team that starts with Copilot and uses the lessons learned to create a blueprint for other workflows and teams; a bank mapping high‑volume work – from customer service to fraud prevention – to prepare for company‑wide transformation and risk management. These are precisely the kinds of approaches that enable organisations, over time, to connect specific solutions into an end‑to‑end chain.

2. From perfect data to “fit for purpose”

For years, the prevailing view was that transformative business value could only be achieved once an organisation possessed a single perfect, structured data layer. Meanwhile, reality has only become messier: structured databases now sit alongside evolving policy documents, emails, PDFs, videos and other unstructured data.

Generative AI’s pattern recognition thrives on this kind of data – endlessly structuring corporate information is therefore not only an impossible task, but also unnecessary. Even cleaning and perfecting data quality often costs more than it delivers. There are countless workflows for which “fit for purpose” data is sufficient – not perfect, but “good enough”.

Of course, there are exceptions. In highly regulated domains, data quality remains crucial. Consider a chatbot that evaluates complex financing applications against policy – here, “good enough” is dangerous. But elsewhere in the workflow, imperfect data can be used without any issues.

It therefore makes sense to shift prioritisation from data‑first to process‑first. Start small: which three to five workflows are valuable enough to design around the data you already have? Instead of having one large central data programme, this creates a targeted approach for each core workflow. Risk and compliance teams play a key role in this journey. They help determine where, for example, 80% accuracy with a human check adds value, and where you can only deploy agents once you can guarantee 99.999% reliability.

3. Closing the digital divide

Although process‑thinking is the foundation for AI transformation within organisations, its success ultimately depends on strong adoption. No process works if employees cannot use the required tools effectively.

Unfortunately, the rise of AI is widening the digital divide. Employees who were already digitally skilled adopt AI faster and more effectively, pulling further ahead of colleagues who struggled with the first digital wave. From a diversity and inclusion perspective, this is a real risk: AI adoption becomes not just a technology issue, but a HR issue as well.

Organisations that want to lead therefore combine three things: they can clearly articulate their strategic ambitions, they work pragmatically from processes to realise those ambitions, and they invest just as much in adoption, reskilling and change capability as they do in technology.

Read the article in AG Connect in Dutch.