The key question in 2026 won’t be what AI can do. It will be where it actually delivers. Five of our experts share their perspective on what really matters in 2026 – and why it’s no longer about chasing the next model, but about focus, context, and the willingness to make decisions.
2026 – a year of new opportunities
2026 will be defined by the demanding climb toward the plateau of productivity. Focus and context will make the difference.
Oliver Macholz, Division Lead Digital Experience
Intelligence without context has no impact
Many companies invest in AI and then wonder why productivity fails to materialize. The problem isn’t the technology - it’s the context.
2025 has made one thing clear: AI does not fail because of insufficient computing power, but because of missing domain and organizational context. Large models are capable of many things — but they do not inherently understand processes, objectives, or decision boundaries. Without proper framing, data remains ineffective.
This is where the next major opportunity lies. Context engineering structures relevant process knowledge so that systems can operate context-aware and support well-founded decisions. Explicit knowledge is formalized and versioned. Implicit knowledge is taken into account wherever it consistently shapes decisions — for example through thresholds, priorities, or exceptions. This creates a reliable decision-making foundation for the effective use of technical intelligence.
Organizations must deliberately design, maintain, and take ownership of context. Only then does technical intelligence translate into real productivity.
Elisabeth Unverricht, Principal Strategy Consultant
Are we measuring AI too soon?
The demand for short-term ROI prevents exactly the work that ultimately makes AI truly valuable.
The demand for measurable returns from AI drives many organizations toward short-term wins — often out of a justified fear of getting stuck in endless pilot loops. But this is where the real risk lies: AI does not fail because it lacks value, but because we force quantification too early.
Technological shifts rarely create value in isolated moments. Their impact emerges gradually — through time savings in knowledge work, improved quality, and new capabilities. These effects are real, yet difficult to capture in traditional KPIs. At the same time, they require precisely the kind of work that looks unattractive in a business case: clean processes, contextually connected data, and a shared understanding of what good outcomes actually look like.
Organizations that reward only short-term results prevent the development of cross-cutting capabilities. The value of AI is already there. We simply need to learn how to recognize it - and measure it better.
Dr. Franz Steinberger, Lead Strategy
Take on the hard things again
Efficiency quick wins rarely deliver on their promise. Real value emerges where things truly start to hurt.
Most prototypes fail shortly after the showcase, because a smooth rollout obscures the fact that real problems can only be solved through sometimes deep interventions in the system. Where internal logics and incentives must be changed, processes rethought, or technical infrastructures integrated or challenged, this is often where true transformational potential lies.
Put differently: meaningful progress only happens where things are potentially difficult and uncomfortable. And this potential is unlocked through a systemic approach to analysis, co-creation with teams when developing measures, and early testing within the organization’s live production environment.
Real progress has never been comfortable. Follow the pain!
Johann Bayerl, Division Lead Digital Communication
The rediscovery of uniqueness
Relevance doesn’t come from more output. It comes from clear differentiation.
Zoe Scaman recently wrote: “Data had its decade. Intuition is making its comeback.” I wouldn’t go quite that far, but one thing is clear: AI makes content infinitely scalable, while attention remains scarce. We all have only 24 hours a day, and more than 80% of digital content generates little to no measurable impact. That’s why more output is rarely the right answer.
Winners don’t respond with more content – they focus on better experiences. They use automation for efficiency, data for relevance, and human creativity and intuition where the difference is truly made.
Not everything needs to be unique. But the things that matter should be.
Dr. Maximilian Franzke, Principal IT Consultant
Rethinking governance
Good governance is not only about prohibitions. It’s about enabling innovative ideas.
Over the past few years, IT governance has become more professionalized across German-speaking regions. Clear rules, centralized platforms, and control mechanisms were necessary. the task is to understand governance not only as a protective mechanism, but as an foundation for impact.
In September, I had the opportunity to explore Africa’s perspective on AI. There, limited AI infrastructure and scarce resources for top-down governance meet enormous potential. AI is not discussed as a threat to existing structures, but as an enabler: AI-amplified community entrepreneurs use it to replace missing teachers, doctors, or experts. Hardy Pemhiwa captures this shift perfectly: "This is not just Africa’s AI moment. This is AI’s Africa moment."
This experience changes my expectations of good governance. What matters is which fundamental problems AI can solve for us – and what we need to do to make that happen.
Glimpse into our insights
Connect