Digital systems decide what AI can do. Denmark’s final letter, Germany’s retreat from fax machines, and Italy’s electronic invoices all point to the same condition: useful AI depends on information that organisations can access, structure, and move through digital channels.
On 30 December 2025, Denmark’s postal service delivered the last letter in its 401-year history. During 2025, the country also removed its remaining 1,500 postboxes from the streets — a physical sign that one information system had reached its endpoint.
Denmark is also ahead in business AI adoption. In 2025, 42% of Danish firms used at least one AI technology, compared with 26% of German firms. Germany stood above the EU average but remained sixteen points behind Denmark, a gap that sits beside a less mature digital infrastructure.
AI adoption follows digitalisation
A 2026 IZA study of 1,907 Danish companies offers a clearer link between the two. The study found that a one-standard-deviation increase in digitalisation was associated with an 8 to 10 percentage point higher probability of adopting core AI systems.
That finding gives the AI agent conversation a less glamorous starting point. Before an agent can act on business information, the information must exist in a usable digital form; otherwise, the agent has little more than scattered records and organisational wishful thinking to work with.
Hemant Kashyap, Chief Product Officer at Kindsight, described the split in executive reactions in an article published on August 21, 2026: “Every leadership team I’ve talked to in the past year arrives at the AI agent conversation from one of two directions: breathless enthusiasm or deep skepticism.”
The Danish study does not settle that argument, but it changes the question. The issue is not only whether an organisation wants AI agents; it is whether its digital systems can support core AI adoption in the first place.
Germany is still replacing the fax
Germany’s healthcare system shows how much basic infrastructure can shape the pace of digital change. In July 2026, the German federal cabinet approved GeDIG, the healthcare data and digital innovation act.
GeDIG phases the fax out of doctors’ offices by autumn 2029 and makes electronic referrals mandatory. Those measures do not sound like the grand finale of technological progress — because they are not — but they establish the digital channels that healthcare data and future AI systems require.
Germany’s 26% business AI adoption rate in 2025 sits in that context. The country is ahead of the EU average, yet its sixteen-point gap with Denmark shows that access to AI technology does not erase the work of digitising everyday operations.
Healthcare provides a particularly stark example because referrals and records must move between offices. A fax-based process and an electronic referral belong to different information systems, and GeDIG sets autumn 2029 as the point by which Germany’s doctors’ offices must leave the former behind.
Digital channels are necessary but not sufficient
Italy offers another version of the same lesson. Its mandatory B2B e-invoicing created a fully digital channel, but overall AI adoption remains below the EU average.
That contrast matters because digitising one process does not automatically digitise an entire business. Italy’s e-invoicing system demonstrates the value of a digital channel, while its AI adoption rate shows that broader adoption still depends on the wider information environment.
Denmark’s postal transition, Germany’s GeDIG timetable, and Italy’s B2B e-invoicing point to infrastructure before spectacle. The hard part of AI adoption may not be persuading companies to discuss agents; it may be replacing the paper, fax, and disconnected records that leave those agents with nothing dependable to use.
The numbers are unromantic, which is often a sign that they matter. Denmark removed 1,500 postboxes, Germany is phasing out fax machines by autumn 2029, and a study of 1,907 Danish companies linked stronger digitalisation with an 8 to 10 percentage point higher probability of adopting core AI systems.
AI agents may still inspire breathless enthusiasm or deep skepticism. Their practical future, however, will be shaped by filing systems, electronic referrals, digital invoices, and the less exciting machinery that lets information move without a stamp or a fax tone.
Based on
