2026 is the year AI must move from experimentation into routine. Many companies are stuck between two poles: on one side they want to "use AI," on the other they don't know where to start. The danger is investing either too broadly (try everything everywhere) or too narrowly (one pilot in one department).
A three-horizons model
We have established a framework that helps leaders set priorities:
Horizon 1: Efficiency (3–6 months) Use cases that make existing work faster or more accurate. Examples: support assistants, document summarization, classification of incoming mail. The ROI is measurable and shows up within a few months. Risk: low. Investment: manageable.
Horizon 2: Growth (6–18 months) Use cases that open new business value. Examples: AI-supported product recommendations, personalized self-service platforms, automated proposal generation. The ROI is indirect (more revenue, higher conversion) and needs a learning phase. Risk: medium. Investment: significant.
Horizon 3: Transformation (18+ months) Use cases that change the business architecture. Examples: fully automated case handling, autonomous supply-chain decisions, radically new customer experiences. The ROI is strategic and hard to pin down on a timeline. Risk: high. Investment: substantial.
The mistakes we observe
Mistake 1: Horizon-3 ambitions without a Horizon-1 foundation. Aiming for autonomous decisions without clean data pipelines is building on sand. Horizon 1 is not just efficiency — it builds the foundation for the later horizons.
Mistake 2: too many Horizon-2 projects in parallel. Every growth project ties up resources for 6+ months. Three parallel initiatives mean 18 months of bound capacity. Focus beats breadth.
Mistake 3: confusing horizon with technology. "We're doing AI now" is not a roadmap. The question is: what exactly do we want to achieve? The technology follows from that, not the other way around.
What is decidable in 2026
Three decisions should be made this year:
Platform decision: which infrastructure carries our AI initiatives over the next three years? The choice between hyperscaler, sovereign cloud, and on-premise has long-term implications. Once made, very expensive to revise.
Data governance: which data assets may be used in which way? Who decides on new use cases? Without clear governance structures, every individual project stalls at data clearance.
Capability strategy: do we build AI capability in-house or work with partners? The answer is rarely black or white — usually a mix. But the weighting must be a deliberate choice.
What can wait
Model selection: the question of whether GPT-5 or Claude 4.6 or an open-source model is "better" is overrated. For 80% of business use cases the model quality is not the bottleneck — the data pipelines, context preparation, and orchestration are.
Marketing communication: talking about AI externally before it works internally is risky. Credibility is tested at the first visible mistake. Deliver first, then communicate.
The next step
An honest readiness assessment is a good starting point. Which Horizon-1 use cases are doable within 6 months? Which data is ready to go? Which governance gaps must be closed before Horizon 2 makes sense?
We support such assessments regularly — typically in 4–6 weeks, with concrete prioritization recommendations at the end. Get in touch if you'd like to learn more.