This research, carried out between 2022 and 2025 by Kevin and Ivo alongside 23 Swiss companies, reveals how generative AI can be effectively scaled across diverse industries such as banking, insurance, healthcare, manufacturing, and consulting. Despite substantial investments in large language models (LLMs), many companies struggle to translate these technologies into broad, strategic advantages. The key to success lies in adopting three core practices: expanding use cases beyond isolated tasks, continuously refining these applications, and promptly discontinuing those that underperform.
A central concept introduced is the ‘AI spine’—a novel organizational structure that integrates domain experts from different business units into a central team managing the AI portfolio. This flexible spine fosters coordination across functions, streamlines processes, and ensures ongoing improvement and compliance. Examples include a Swiss bank that significantly boosted customer service efficiency with generative AI tools, and a medical coding company that not only cut operational costs but also created new revenue streams by commercializing AI-driven coding solutions.
The AI spine is financed independently and overseen by C-suite executives, ensuring strategic alignment and agility in funding high-impact projects. It contrasts with traditional hub-and-spoke models by centralizing cross-unit business knowledge rather than dispersing technologists across units. This structure also embeds continuous feedback loops involving business owners, tech specialists, compliance experts, and end users to evolve solutions responsive to real-world needs.
While alternatives like GenAI units and squads offer some benefits, they often fall short of achieving enterprise-wide integration and sustained innovation. The research highlights the necessity for top-level commitment, cross-unit collaboration, clear metrics for value, and ongoing dialogue between technology and business stakeholders. Leaders seeking transformative impacts with generative AI must architect such collaborative frameworks to unlock and sustain economic value at scale.
