Analysis ·
Why enterprise AI is stalling, and what it takes to move beyond the pilot
Original title, in French: “Pourquoi l’IA d’entreprise stagne, et ce qu’il faut pour passer au-delà du pilote”
- Publication
- Unite.AI (United States)
- Author
- Unsigned
- Published on
- 22 September 2026 · read in the page metadata · updated on 22 September 2026
Summary
The article analyses why enterprise AI projects so often remain stuck at the pilot stage. According to McKinsey & Company's State of AI, nearly two-thirds of organisations have not yet begun to deploy AI at enterprise scale. According to IDC, 88% of AI proofs of concept never reach production. The author explains that the real obstacle is not technological capability but integration into business processes, change management and governance. He distinguishes between visible AI agents (assistants) and invisible agents embedded in workflows, and recommends right-sizing the technical architecture: for 80 to 90% of use cases, a small model is enough, rather than a frontier model.
Summary based on the full text of the article.
Read the article on Unite.AI →Figures cited
-
66 %
share of organisations that have not begun deploying AI at enterprise scale
companies surveyed, worldwide
« près de deux tiers des répondants déclarent que leurs organisations n’ont pas encore commencé à déployer l’IA à l’échelle de l’entreprise. »
Source: McKinsey & Company, State of AI
-
88 %
share of AI proofs of concept that do not reach production deployment
companies, worldwide
« Les recherches d’IDC ont montré que 88 % des preuves de concept d’IA n’atteignent pas le déploiement en production, ce qui signifie que sur 33 lancées, seules quatre aboutissent. »
Source: IDC
-
85 %
share of enterprise use cases that do not require a frontier AI model
enterprise use cases, estimate
« pour 80 à 90 % des cas d’usage en entreprise, un modèle de pointe est inutile. »
Classification
- Country
- Worldwide scope
- Company size
- All sizes
- Industries
- All industries, or industry not specified
- Companies
- McKinsey & Company (study author), IDC (study author)
- Keywords
- scaling, proof of concept, AI agents, governance, business transformation, small language models
Entry written by Marsanne Studio from the cited article, with the help of an AI model (claude-sonnet-5). The rights to the article and its illustrations belong to Unite.AI. Entry translated from French by Claude Sonnet 5.5.