AI strategy
AI only works where strategy and data are in order.
Yoline’s core statement: an AI strategy without data architecture is activity without a system.
AI strategy as the basis
Most AI projects begin with tool selection: ChatGPT, Copilot, Claude. Yoline begins with strategy: where does AI actually have an effect? Which use cases have a data basis? Which impact should become measurable? Tools follow.
Three patterns that repeat
First: tools are introduced before the use case is named. Licences are paid for and hardly used. Second: the data basis is missing. If data are scattered, incomplete or dirty, AI produces hallucinations. Third: the profitability calculation is missing. What does the introduction cost compared with the real benefit, and over what period does it pay for itself?
Five fields of work
Use case prioritisation
Which AI applications have the greatest impact per effort? Economic evaluation as the selection logic.
Data requirements
For each use case: which data in which quality. Gaps named explicitly.
Tool selection
Strategic, by use case. Generic chat LLMs and Copilot tools versus specialised solutions. Pragmatic.
Implementation
Operationally with Désirée Frei. Integration into existing systems.
Measuring impact
What is the KPI? How is it collected? Who evaluates? Steering as a mode.
Marco shares his know-how
Marco shares his know-how at Digitourism as a speaker and lecturer. 18 Insights articles on AI strategy and AI practice document the method. The consulting approach grows out of teaching and engagement practice.
AI needs a basis
AI strategy only works with a data and CRM basis. Yoline integrates the three disciplines as a connected system. AI visibility (GEO) is a sub-discipline (see the GEO optimisation service).
Platform experience
Marco teaches AI strategy as a speaker and lecturer at universities. This teaching connects method and engagement practice.
Architecture sprint: on request, 4 to 8 weeks. Engagement: on request, always after the sprint. Initial conversation of 30 minutes, free of charge.
What does not work
A ChatGPT licence for everyone and hoping something happens. Choosing tools without a use case. Data cleansing as an afterthought. Missing governance for generative AI in customer contact. And the biggest mistake: introducing AI as an IT project. The strategic decision is missing.
How does AI strategy differ from AI consulting?
Strategy is architecture and a roadmap over quarters. Consulting addresses one specific question at a single point. Yoline works primarily at a strategic level, but can also provide half-day sparring on specific AI topics.
Do I need technical AI knowledge?
No. Marco works at strategy level with management. Désirée Frei handles the technical depth operationally. You need the authority to decide and an honest view of the business reality.
Which AI tools does Yoline cover?
Tool-neutral. Yoline works with the leading chat LLMs, Copilot tools and custom LLM set-ups. The selection follows the use case and the data architecture.