When AI fails, it's not only the model. It's data, rights and workflows too. Three figures from current studies show where it gets stuck.
The results do not come
Only 25% of AI initiatives delivered the expected ROI, and only 16% scaled enterprise-wide (IBM CEO study, 2,000 CEOs).
The data is not ready
Through 2026, Gartner expects organisations to abandon 60% of AI projects that lack AI-ready data.
The rights are not settled
Only 21% say they have a mature governance model for agentic AI, while 74% expect to use agents by 2027 (Deloitte). Without such rules, an agent sees too much or may do nothing at all.
What AI needs
demi is built in three layers, in exactly this order: the Data Layer brings the data from every system together, and every value keeps its source. The Permission Layer turns the org chart into permissions. Only then does AI work in the Intelligence Layer, exact and traceable.
Sources
- IBM Institute for Business Value, 06/05/2025: IBM Study: CEOs Double Down on AI While Navigating Enterprise Hurdles
- Gartner, 26/02/2025: Lack of AI-Ready Data Puts AI Projects at Risk (Q&A with Roxane Edjlali)
- Deloitte, 24/04/2026: Business and IT leaders report AI agents are scaling faster than their guardrails
