Ask almost any executive team today whether artificial intelligence matters to their business and the answer will be yes. Ask which AI initiatives are already generating measurable, scalable business value and the conversation often becomes more difficult.
This is the paradox of the current phase of AI adoption. The technology is advancing at extraordinary speed. Employees are experimenting with generative AI, copilots are becoming part of standard software packages, and new use cases emerge almost every week.
Yet many organisations are still struggling to move from experimentation to impact.
The bottleneck is increasingly not technology. It is execution.
“The question is no longer whether AI can create value. The question is how quickly organisations can identify where it creates value - and scale it.”
Frank Dannacher
That observation was one of the starting points for the Swiss AI Playbook, commissioned by digitalswitzerland and developed by Implement Consulting Group. Its objective is deliberately practical: rather than producing another report about what AI might eventually do, it provides leaders with guidance on what they can do now.
Working with large and mid-size companies operating globally, our experience suggests that four questions matter in particular.
1. Use Case Prioritisation: Where does AI actually create Value?
There is no shortage of AI ideas. Employees, vendors and management teams can quickly generate dozens or even hundreds of potential applications.
The challenge is prioritisation.
A useful AI use case needs to solve a real business problem. That means looking beyond technological fascination and asking:
What problem are we trying to solve? Who benefits? What measurable value could be created? And can the solution realistically be implemented and scaled?
Instead of asking, “Where can we use generative AI?”, a sales organisation might ask, “Why do our account managers spend several hours preparing for client meetings?”
Instead of asking, “Could we build an AI chatbot?”, a service organisation might ask, “Why does it take our employees so long to find the right answer to a customer question?”
Starting with the business problem rather than the technology makes it much easier to distinguish valuable applications from interesting demonstrations.
The principle is simple: start small, learn quickly and scale selectively.
A focused experiment delivering tangible value within weeks can be more useful than an ambitious enterprise-wide AI programme spending months designing the perfect future architecture.
2. Is the Organisation technically ready without Overengineering?
Once promising use cases have been identified, organisations need the data and technology foundation to support them.
This is where companies can fall into two traps:
- The first is underestimating the foundation. AI applications depend on accessible, reliable and appropriately governed data. A powerful model cannot indefinitely compensate for fragmented information, unclear ownership or poor-quality source data.
- The second is overengineering the foundation before creating value.
Not every AI application requires a completely new enterprise data platform. Modern SaaS solutions, existing cloud environments and increasingly capable off-the-shelf AI tools allow organisations to test many use cases with relatively lightweight architectures.
The right question is therefore not:
“What is our ultimate AI architecture?”
but:
“What is the minimum viable technical foundation for this use case and what will we need if we scale it?”
This is particularly relevant for SMEs. AI is not a game reserved for companies with enormous technology budgets and large data science departments.
Competitive advantage may increasingly come not from owning the most sophisticated AI technology, but from applying widely available technology better than competitors.
3. Are People actually using it?
Technology alone creates no productivity improvement.
People using technology differently do.
A technically excellent AI solution that employees do not trust, understand or integrate into their daily work creates little value. Conversely, relatively simple tools can generate significant impact when embedded in everyday workflows.
AI adoption therefore requires more than conventional software training.
Employees need to understand what AI is good at, where its limitations lie, how to formulate effective prompts, how to evaluate outputs critically and when human judgement remains essential.
But organisations also need to create space for experimentation.
Some of the most valuable applications will not be invented centrally. They will emerge when people who deeply understand a business process discover that AI allows them to perform part of that process fundamentally differently.
The role of the organisation is therefore not only to deploy AI solutions. It is to build AI capability - combining clear guardrails with curiosity and empowerment.
4. Is Leadership treating AI as a Transformation?
This may ultimately be the most important question.
AI is often delegated to IT, Digital or an innovation team. Those functions have important roles to play. But if AI changes processes, roles, decisions and potentially even business models, it cannot remain a technology initiative.
It becomes a management transformation.
“AI may be powered by technology, but capturing its value is a transformation challenge. And transformation is ultimately about changing how people work, decide and collaborate.”
Markus Koch
Leadership therefore needs to provide direction without pretending to know exactly where the technology will lead.
That requires an unusual combination of ambition and pragmatism. Leaders need to articulate where they expect AI to create value while accepting that experiments will fail, technologies will change and today's preferred solution may look very different twelve months from now.
Traditional transformation programmes often try to reduce uncertainty before implementation begins.
AI transformation requires organisations to execute while uncertainty remains high.
That changes the role of leadership. Rather than designing every detail centrally, leaders need to establish priorities, create simple governance, remove bottlenecks and accelerate learning.
And importantly, they need to use the technology themselves.
Employees will quickly notice the difference between executives who talk about AI and executives who have begun changing their own way of working with it.
The Swiss AI Playbook: From AI Strategy to an Execution Engine
Taken together, these observations point to a broader shift. During the first phase of generative AI, the central question was: What can this technology do?
The next phase is about something different:
How quickly can organisations turn what is technologically possible into repeatable business value?
That requires three capabilities working together:
- Use cases
that solve meaningful business problems and create measurable value. - Data and platforms
that provide a pragmatic but scalable technical foundation. - People
who understand, trust and effectively use AI.
Surrounding all three is the fourth dimension: leadership and transformation.
Weakness in any one of these areas can prevent scaling. A company may have brilliant use cases but insufficient data. It may have excellent technology but poor adoption. Or it may have enthusiastic employees running dozens of experiments without strategic prioritisation.
Companies therefore do not simply need an AI strategy.
They need an AI execution engine: a repeatable way of identifying opportunities, testing them rapidly, measuring their impact and scaling the ones that work.
The backbone can be surprisingly simple:
Identify → Prioritise → Experiment → Measure → Scale
Then repeat.
Over time, the organisation becomes better not only at using particular AI tools but at absorbing continuous technological change. That capability may ultimately prove more valuable than any individual AI application.
The real AI Race is an Execution Race
Switzerland starts from a strong position: world-class research, highly skilled employees, innovative companies and a business environment built around quality and trust.
But AI technologies are increasingly accessible globally. The models available to a Swiss SME are often the same models available to competitors elsewhere. Differentiation therefore comes from how effectively organisations apply them.
This is why the Swiss AI Playbook focuses deliberately on implementation rather than prediction. It provides practical frameworks, examples and checklists that help organisations take the next step now. Because waiting for AI to stabilise is probably not a viable strategy.
There will always be a better model, a new platform or another technological breakthrough around the corner. The companies that develop the ability to experiment, learn and scale continuously will have an advantage over those waiting for certainty.
AI is often described as a technology race. For most companies, that is the wrong race.
“For most companies, the AI race is not about building better technology. It is about becoming better at turning technology into business impact.”
Markus Koch
Companies are unlikely to build the world's best foundation model. Nor do they need to. Their challenge is to become better than their competitors at turning AI into improved decisions, better customer experiences, more productive processes and new sources of growth.
So perhaps the most useful question for management teams is no longer:
“Do we have an AI strategy?”
It is:
“How good are we at turning AI into business impact?”
That is the question the Swiss AI Playbook is designed to help organisations answer.
About the Swiss AI Playbook
The Swiss AI Playbook was commissioned by digitalswitzerland and developed by Implement Consulting Group. It provides practical guidance, frameworks, examples and checklists for organisations seeking to move from AI experimentation to scalable business impact. The playbook addresses four interconnected dimensions: use cases, data & platforms, people, and leadership & transformation.
About the Authors
Frank Dannacher is a Partner at Implement Consulting Group and one of the driving forces behind the Swiss AI Playbook. He advises organisations on digital and AI transformation and on turning emerging technologies into tangible business impact.
Markus Koch is a Partner at Implement Consulting Group specialising in transformation and strategy execution. He supports leadership teams in turning strategic ambitions, including the opportunities created by AI, into organisational change and sustainable business results.