If you look at the daily headlines, it is easy to get caught up in the AI arms race. But I think we need to shift our focus. Instead of obsessing over which company has the most advanced foundation model, we should judge AI adoption through a much simpler, more practical lens: productivity gains.

We are finally at a point where AI tools are delivering on this promise. Take a tool like Claude Code, for example. It is making it possible for far more people to develop high-quality code at an unprecedented speed. But as these productivity gains accelerate, they are inevitably going to create new bottlenecks. If AI can write a feature in minutes, we suddenly hit a wall if we still insist on traditional human code reviews. It also creates structural challenges for the industry: how do junior developers enter the job market and learn the ropes when AI has essentially automated the entry-level work?

These are tough questions, but they are the right questions to be asking. It means the technology is actually working.

The European Perspective

‍
Here in Europe, there is often a tendency to wring our hands about the fact that we don't have the same amount of frontier AI labs as the US. But a recent study from CEPR (Centre for Economic Policy Research) analysing over 12,000 European firms paints a very encouraging picture.

The research shows that we don't necessarily need to build the models to harvest the gains. According to the data, AI adoption increases labor productivity levels by 4% on average across the EU—and importantly, with no evidence of reduced employment in the short run.

But there is a catch.

The Uneven Distribution of Gains

‍
That 4% is just an average. The CEPR study highlights a crucial point: these productivity gains are highly uneven. Medium and large firms are the ones currently experiencing the strongest productivity boosts.

Why? Because they have the resources to invest in what the researchers call "intangible assets and human capital". In plain English: investments in software integration and employee training act as massive productivity multipliers. If you just buy an AI license and hope for the best, you get very little. If you invest in the right software and train your people to use it, productivity spikes.

Leveling the Playing Field

‍
This uneven distribution is a problem. If only large enterprises have the budget and capacity to properly integrate AI, smaller companies will be left behind in the productivity race.

This is exactly what we are trying to solve at Raffle. We want to help smaller and mid-sized organisations realise these exact same AI productivity gains without needing a massive enterprise IT or training budget.

We do this by delivering search and chat tools that handle the heavy lifting of information-related tasks. Think about municipalities, legal departments, or any rule-dense area where employees spend a massive chunk of their week just trying to find the right internal guidelines, policies, or case files. By applying AI to solve the retrieval problem, we turn an expensive, time-consuming bottleneck into a fast, automated process.

Ultimately, AI adoption shouldn't be about having the shiniest new tech. It is about removing the friction in our daily work so we can get more done.

‍