According to a recent study by the Swiss Federal Statistical Office, the Swiss have a clear edge. Around 47% of 16- to 74-year-olds use AI—well above the European average of 33%. The gap is even wider in professional settings: 73% of Swiss respondents said they use AI at work, compared with 46% in the EU.

AI, then, has “arrived” in Switzerland. But what do these figures mean in practice—say, for communications departments? The Commtech Index Report, which surveys more than 500 communications teams across the German-speaking world each year, offers a more nuanced picture in its 2025/26 edition. It identifies an emerging digital divide, with large firms pulling ahead of smaller ones at pace. It also finds that resources are flowing into areas with quick, visible returns rather than longer-term structural projects. “Data are the real obstacle to digitalising communications—not AI,” the report concludes.

This is hardly surprising. The truly significant productivity gains require agentic AI, and the path there demands that organisations tackle formidable barriers: consolidating data silos, and reformatting information so that machines can actually use it. This is anything but trivial. Unsurprisingly, 72% of Swiss participants in the Commtech survey identify systems integration as the single biggest hurdle to deploying the new technology.

Teething troubles with agentic AI are widespread. As a McKinsey study showed earlier this year, the share of companies that have deliberately deployed AI agents and restructured their workflows around them remains in the low single digits.

aus: The state of AI in 2025: Agents, innovation, and transformation, McKinsey, November 2025

How, then, does a communications or marketing department make AI work? Despite appearances, this is not – at least at the outset – a task for the IT department. It belongs to the people who stand to benefit. A strategic approach pays dividends: define objectives carefully, negotiate priorities, and conduct a thorough audit with clear definitions.

A few questions worth asking at the outset:

• What are our primary objectives?
• Which channels – website, newsletters, social media, press, events – do we actively manage, and what workflows does each entail?
• What knowledge bases and templates do we rely on regularly?
• How do we manage quality?
• Is information stored somewhere accessible to everyone responsible for it?
• Who owns the content, quality and maintenance of the underlying data?

That last point – data stewardship – is especially critical in a technological context. An AI system is only as good as the database it draws on. This need not mean hiring someone new. But it does require visible accountability within existing roles: whoever owns content must also own the data behind it. Datengrundlage dahinter.

Once the audit is complete and responsibilities are clear, implementation can begin. The question then becomes: which problems or tasks could agents most usefully address? Where is the bottleneck? In media monitoring? Trend-scouting? Content production? The final push before a press conference? Knowledge management? Reporting? The answer differs in every organisation – and that, precisely, is the point. AI does not fulfil its potential in the abstract; it does so where it meets well-defined processes and well-maintained data.

AI tools are not the first step into digital communications – they come last. Those who do the groundwork well stand a realistic chance of getting what the technology promises: greater productivity and relief from the grind of routine tasks.