7% Intensive AI Users Expose EU's Productivity Gap

On Wednesday (June 24), the European Central Bank put a hard number on Europe’s AI gap: 70% of surveyed firms use AI, but only 7% of euro area firms are intensive AI users.

That timing matters because the debate has moved past whether companies have tried AI. The sharper question is whether they have made it part of daily operations. The new ECB research, reported by PYMNTS, suggests Europe has broad AI exposure but shallow operational depth.

June 24 ECB data turns Europe’s AI story from adoption to intensity

The headline number sounds strong. If 70% of surveyed firms are using AI, Europe is not sitting out the technology cycle. But the ECB’s deeper finding cuts through that optimism: most firms are using AI only infrequently or moderately.

“Nearly half of the firms that were not using AI in 2025 plan to invest in it in 2026,” ECB wrote on its blog. “At the same time, most firms report using AI only infrequently or moderately, with only 7% of euro area firms reporting intensive use.”

That 7% intensive AI users figure is the real signal. It says the euro area’s AI challenge is not awareness. It is conversion. Companies are testing tools, buying licences, and experimenting with use cases, but far fewer are reorganizing work around AI.

XOOMAR analysis: this distinction matters because adoption statistics can flatter the market. A company that uses AI for occasional drafting or internal search counts as an AI user. That does not mean AI has changed its cost structure, service model, product cycle, compliance process, or decision-making cadence.

After the 70% headline, the 7% intensive AI users figure does the damage

The ECB defines intensive AI use as more than infrequent or moderate usage. That makes the euro area split look stark.

The sector detail is important. The ECB said intensive AI use is especially prevalent in services, particularly in “high-tech, knowledge-intensive” services such as information and communication.

“These sectors include developers and providers of AI tools, which tend to be highly digitalized,” the ECB blog said. “They also have access to abundant data and computing infrastructure and employ workers with strong technical skills.”

That is not a small qualifier. It suggests intensive AI adoption depends less on enthusiasm and more on inputs: data, infrastructure, technical staff, and processes that can absorb model output.

XOOMAR analysis: if those inputs are missing, AI stays at the edge of the company. It helps with tasks. It does not become a production system.

The immediate read: experiments are easy, workflow rewiring is not

The ECB data points to a management problem disguised as a technology problem.

Light AI use is easy to start. A team can test a chatbot, draft a memo, summarize documents, or automate a small internal workflow. Intensive use is harder because it asks companies to change how work is approved, measured, staffed, and governed.

That shift can touch:

• Data readiness: AI systems need usable, accessible, and trusted data.
• Digital infrastructure: Older systems can slow integration.
• Staff skills: Firms need workers who can translate AI output into business processes.
• Governance: Managers need rules for review, accountability, and vendor oversight.
• Budget discipline: AI spending has to move beyond licences into integration and restructuring.

The ECB makes the cost issue explicit. It found that more than 84% of companies reporting intensive AI use have invested in the technology, compared with 34% of moderate users. Nearly all intensive users, 99%, planned to invest in AI this year, setting aside roughly 20% of total investment for AI-related activities.

“This shows that investments that go beyond purchasing licences for general AI tools typically require more substantial funding,” the ECB wrote. “Indeed, integrating AI into core processes, such as developing customised solutions or upgrading digital infrastructure, often entails larger and longer-term restructuring.”

That sentence explains why the 7% share is low. Intensive AI is not just procurement. It is restructuring.

The 2026 budget signal separates serious users from casual users

The most useful part of the ECB research is not the current adoption rate. It is the investment split between intensive and moderate users.

A firm that spends meaningfully on AI-related activities is making a different bet from one that buys general tools for staff. The first is trying to build repeatable capability. The second may still be testing whether AI belongs in the workflow at all.

Executives, workers, regulators, and investors will read the same 7% intensive AI users figure differently.

• Executives: The appeal is clear, but the ECB numbers imply that deeper use requires larger and longer-term restructuring. That raises the bar for return on investment.
• Workers: PYMNTS cited separate research saying there is “no single playbook for integrating AI into the workforce.” Firms are hiring AI expertise, reconfiguring workflows, or doing both.
• Regulators and policymakers: Low intensive adoption may mean fewer immediate deployment risks, but it also raises a competitiveness question if deep usage stays concentrated in a small group of firms.
• Investors: XOOMAR analysis: the firms that can show AI is embedded in core processes will be easier to distinguish from those relying on AI language in strategy decks.

For readers tracking the broader AI stack, XOOMAR has covered adjacent parts of the market in Baseten Funding Frenzy Tests a $13 Billion AI Wager and consumer-facing AI in Buried Apple Intelligence Features Rescue iPhone AI. The ECB data adds the buyer-side lens: are companies actually turning AI into routine operating capacity?

Separate EU datasets show adoption varies, but don’t blur the scopes

The ECB findings refer to surveyed euro area firms. That distinction matters because the article title frames the issue around EU companies, while the key intensive-use figure comes from the ECB’s euro area research.

A separate Euronews report, citing European data, said the use of at least one type of AI technology among European businesses stood at around 19.95% on average, with large country gaps. It also reported that AI use among European enterprises had grown by 12.30% since 2021. Those figures are useful context, but they are not the same dataset as the ECB survey.

Keep the scopes clean:

XOOMAR analysis: the consistent message across these sources is not that Europe lacks AI activity. It is that activity is uneven, and depth is the scarce commodity.

The next decision point is proof, not pilots

The next phase of Europe’s AI market will not be decided by how many firms have tested generative tools. It will be decided by how many can turn those tools into repeatable, governed, measurable business processes.

The ECB’s investment numbers give readers practical signals to watch in 2026:

• Dedicated AI budgets beyond software licences.
• Infrastructure upgrades tied to AI deployment.
• Custom AI solutions rather than generic tool use only.
• Workforce redesign that changes how tasks are allocated.
• Governance processes for model use, review, and accountability.
• Measured outcomes that show AI has moved from experiment to operating system.

The thesis is simple: Europe’s AI footprint is already wide, but its productivity engine is still shallow. Evidence that would confirm improvement is a rising share of intensive AI users, paired with higher investment in integration and core-process redesign. Evidence that would weaken it is continued growth in casual AI use without a matching rise in intensive deployment.

Impact Analysis

• Europe’s AI gap appears to be about operational depth, not basic adoption.
• The 7% intensive-use figure suggests many firms have yet to turn AI into productivity gains.
• Future competitiveness may depend on whether companies move from experimentation to daily AI-driven workflows.