The trend radar shows a shift in the strategic centre of gravity. The most striking individual change is AI Strategy, which moves directly from the “Create” zone into the “Act” zone. According to the study, the challenge is no longer getting started, but scaling: linking AI initiatives to measurable business results. For mid-sized companies, this challenge becomes most tangible where technology has a direct impact on measurable outcomes—for example, across the customer journey in marketing, sales and service. This is where it becomes clear fastest whether an initiative contributes to revenue, conversions and customer retention, or has no lasting impact.
More important than what’s at the top, however, is the foundation. The radar’s conclusion is clear: strategic relevance is shifting away from isolated innovation topics and towards the deeper prerequisites for an organisation’s ability to act. Data quality, data protection, reliable operational processes and a learning organisation are no longer supporting issues; they are the conditions that make everything else work.
AI rarely fails because of the model in mid-sized companies; it almost always fails because of the data. The trend radar backs this up with numbers.
-
67 percent of organisations do not fully trust their own data when making decisions.
-
52 percent of surveyed companies cite data quality and availability as the biggest obstacle to AI adoption.
This explains why so many AI pilot projects get stuck in the pilot phase. In many companies, each department has worked with its own system for years. This has led to silos in marketing, sales and service. Sales cannot see marketing data; when a new customer is onboarded, information is entered again; and knowledge about the customer gets lost along the way. The result is no 360-degree view of the customer, simply because the data is stored in separate systems.
The radar names the real task directly: Data governance is not primarily a technical issue, but an organisational one. Who maintains which data, who grants access to it, and who is responsible for it? Our answer is an architectural hub. A Customer Platform such as HubSpot brings customer data from marketing, sales and service together in a shared data model. When the ERP system is also connected, transactional data such as orders, invoices and delivery status flows into the customer record.
The effect is clear: everyone who interacts with customers can see all relevant information in the CRM, and AI can access the same complete context. This unified customer record is essential for AI to work with reliable data, because otherwise every model encounters gaps and contradictions. The best-of-breed approach remains valid, but it needs a core made up of an ERP system and a Customer Platform. More specialised systems, such as industry-specific solutions, e-commerce platforms or BI tools, are then connected to this core in a targeted way.
The customer journey is becoming both more automated and more human. This is not a contradiction, but the central tension of the coming years.
On the one hand, routine interactions are disappearing into automation. On the other, the radar shows a growing demand in the opposite direction, with the new field of action called Human-to-Human Experience.
-
According to the cited Gartner survey, 64 percent of customers would prefer companies not to use AI in customer service.
-
53 percent would even switch providers if they knew about it.
-
For Germany, the cited trend monitor confirms that 75 percent still consider interpersonal contact in customer service important.
In marketing, automation handles segmentation, campaign logic and lead qualification. People shape positioning and messaging. In sales, the CRM provides context, the next best action and smooth handovers. People lead the conversations that build trust. In service, bots and self-service handle recurring enquiries. People step in when cases are complex, emotional or escalated—exactly where customer loyalty is won or lost.
The key point is that this division of labour only works with a shared data foundation. If a service employee handling a sensitive escalation cannot see what marketing promised and sales committed to, the human interaction may be well-intentioned but lacks context. The Human-to-Human Experience depends on data architecture; it is not opposed to it.
What Role Do AI Agents Play in the Value Chain?
AI agents are shifting the boundary between assistance and semi-autonomous execution. They no longer just respond to individual requests; they plan and carry out multi-step tasks, independently accessing data and systems along the way. The idea of agents taking over entire process chains is no longer science fiction: they can handle everything from preparing a quote and confirming an order to creating an invoice.
For mid-sized companies, this means two things:
-
First, real leverage comes when an agent orchestrates an entire workflow—not when it saves a single click. An agent that takes a quotation process from the initial enquiry through data enrichment to a ready-to-send draft changes how value is created. An agent that merely suggests a text snippet is a convenience.
-
Second, data quality determines how useful an agent can be—and matters more than in any previous application. An agent operating on inconsistent data will make incorrect decisions independently and more frequently. This brings us back to architecture: only a consistent customer data model makes agents reliable across marketing, sales and service.
For an agent to help rather than cause harm, it needs clear guardrails. Companies need to answer a few fundamental questions:
-
What tasks do we delegate to AI?
-
What data is AI allowed to use?
-
What rules must AI follow?
-
Where is oversight needed—for example, through a deliberate human-in-the-loop process in which a person always makes the final decision?
-
How do we shape collaboration between people and AI so that it is productive and responsible?
An agent that operates without such rules and human oversight is not progress, but a risk. The study rightly warns against two patterns: blindly trusting AI suggestions and being too quick to question our own judgement as soon as a machine provides an answer.
As soon as a company collects customer data, clear rules apply. You need to know what data you hold about a person and why; for many uses, you need their consent, for example for marketing emails; and you must be able to provide or delete the data upon request. If AI is used in customer interactions, you must also be open about the fact that a chatbot is a machine, not a person.
These rules are set out in the revised Swiss Data Protection Act. Companies that also serve customers in the EU must comply with the local requirements, which are somewhat stricter. For a Swiss SME, this is no cause for concern, but it is a reason to get things right from the start.
The good news: A central platform for all customer data takes care of much of this. When all the data is held in one system rather than scattered across disconnected solutions, it is clear who is responsible for each type of data, every consent is recorded and traceable, and data can be properly deleted. This makes data protection a feature of the system rather than a laborious manual task. The architectural hub pays off a second time: first for AI, and now for data protection.
Even the best architecture will not transform an organisation on its own. The trend radar is clear: new systems often fail not because of the technology, but because they run up against established habits and power structures without leaders actively investing in change. This is exactly why a learning organisation is a foundation: every CRM, digitalisation or AI initiative is also a change initiative.
For mid-sized companies, this means thinking about the data model, processes and adoption together. Introducing technology alone creates a digital shell around old structures. Involving employees early, clarifying responsibilities and building skills can turn a system into genuinely new ways of working. That is why embedding an implementation in the organisation must be part of the project, not an afterthought to be addressed separately later.
Responsibility is distributed, but it converges around a shared architecture. Three roles, three priorities, one foundation.
Conclusion
The 2026 trend radar is encouraging. It shows that major topics such as AI, digitalisation and customer experience are not separate large-scale projects that all have to be tackled at once. They rest on a shared foundation of clean data, clear processes and clear responsibilities. And every company has this foundation in its own hands.
That is the real good news for mid-sized companies: they do not have to tackle all 15 areas. They need to get one thing right: the architectural hub that brings together data, CRM and an end-to-end Customer Journey. Investing in this means building the foundation for AI, digitalisation, data protection and a genuine customer experience to work together.
Mid-sized companies are at an advantage here: shorter lines of communication, faster decisions and closer customer relationships. Laying the foundation now creates a lead that grows with every step. Get one thing right, and many things become possible.
