--- title: "AI adoption in companies: only 7% achieve measurable value and 23% scale up with moderate returns — Marsanne Studio AI watch" description: "According to a TCS/Fast Company study of 380 executives, only 7% of companies derive measurable value from AI." url: "https://marsanne.studio/en/ai-watch/ai-adoption-in-companies-only-7-achieve-measurable-value-and-23-scale-0f4e29/" lang: "en" --- [Marsanne Studio](https://marsanne.studio/en/) [← Back to home](https://marsanne.studio/en/ai-watch) [← The AI watch](https://marsanne.studio/en/ai-watch) Data study · 22 September 2026 # AI adoption in companies: only 7% achieve measurable value and 23% scale up with moderate returns Original title, in Spanish: “Adopción de la IA en las empresas: solo el 7% logra un valor medible y el 23% escala con retornos moderados” **Publication** Merca2 (Spain) **Author** Inversión **Published on** 22 September 2026 · read in the page structured data ## Summary A report by TCS and Fast Company, based on a survey of 380 senior executives across a range of industries, shows that only 7% of companies derive measurable value from AI. 23% achieve moderate returns by scaling up, 23% remain stuck at the pilot stage, 21% are in early experimentation, 18% have not yet adopted AI and 8% have abandoned their initiatives. Return on investment is the main obstacle for 39% of companies. The report distinguishes "AI native" companies, which completely redesign their processes, from the others, which merely add AI to existing processes. Examples are given: Mastercard improved its fraud detection by 300% by giving all its employees access to Copilot, and Autodesk invested $350 million over three years in an AI Academy. The article also links these findings to the Spanish SME landscape. Summary based on the full text of the article. [Read the article on Merca2 →](https://www.merca2.es/2026/09/22/adopcion-ia-empresas-valor-medible-2459136/) ## Figures cited - 7 % Share of companies achieving measurable value from AI 380 senior executives, various industries > « Solo un 7% declara valor medible gracias a la IA. » Source: Informe de TCS y Fast Company - 23 % Share of companies scaling up AI with moderate returns 380 senior executives, various industries > « El 23% escala con retornos moderados » Source: Informe de TCS y Fast Company - 23 % Share of companies stuck at pilots that fail to take off 380 senior executives, various industries > « otro 23% se queda en pilotos que no despegan » Source: Informe de TCS y Fast Company - 21 % Share of companies in the early experimentation phase 380 senior executives, various industries > « el 21% sigue en fase de experimentación temprana » Source: Informe de TCS y Fast Company - 18 % Share of companies that have not yet started using AI 380 senior executives, various industries > « un 18% ni siquiera la ha tocado » Source: Informe de TCS y Fast Company - 8 % Share of companies that have paused or abandoned their AI initiatives 380 senior executives, various industries > « Un 8% llegó a pausar o abandonar sus iniciativas después de haberlas arrancado. » Source: Informe de TCS y Fast Company - 39 % Share of companies for which return on investment is the main obstacle 380 senior executives, various industries > « El retorno de inversión es el principal obstáculo para el 39% de las compañías, según el informe. » Source: Informe de TCS y Fast Company - 70 % Share of leading companies running several critical AI systems in production Leading companies in the sample > « El 70% de las empresas líderes opera con varios sistemas de IA críticos en producción » Source: Informe de TCS y Fast Company - 13 % Share of companies with a critical system in which AI is integral to the business 380 senior executives, various industries > « solo el 13% de las encuestadas reconoce tener un sistema crítico donde la IA es integral para el negocio » Source: Informe de TCS y Fast Company - 300 % Improvement in Mastercard's fraud detection through AI Mastercard > « Una mejora del 300% en la detección de operaciones fraudulentas. » - 26 % Share of companies without a human-AI collaboration model 380 senior executives, various industries > « el 26% de las empresas no cuenta con ninguno » Source: Informe de TCS y Fast Company - 39 % Share of all companies giving most or all employees access to AI 380 senior executives, various industries > « más de la mitad de las AI-native dan acceso a la IA a la mayoría o a todos sus empleados, frente al 39% del total » Source: Informe de TCS y Fast Company - 28 % Share of companies citing lack of in-house skills as the main barrier 380 senior executives, various industries > « el 28% señala la falta de habilidades internas como su mayor barrera » Source: Informe de TCS y Fast Company - 36 % Share of all companies for which data quality or governance is the main obstacle 380 senior executives, various industries > « Para el 30% de los líderes y el 36% del conjunto, la calidad, el gobierno o la infraestructura de datos es el mayor freno. » Source: Informe de TCS y Fast Company - 350,000,000 USD Autodesk's investment in its AI Academy over three years Autodesk, over three years > « Autodesk puso en marcha una AI Academy de un año con 350 millones de dólares comprometidos a tres años para formar en IA » - 82 % Share of leading companies measuring AI-related revenue growth Leading companies in the sample > « el 82% mide crecimiento de ingresos de productos o servicios con IA, frente al 51% del total » Source: Informe de TCS y Fast Company ## Classification **Country** Spain **Company size** All sizes **Industries** Banking, insurance, asset management, finance departments · Software vendors, IT services companies, telecoms · All industries, or industry not specified **Companies** Mastercard (deploys AI), Autodesk (deploys AI), Canva (cited), TCS (study author), Fast Company (study author), HP (deploys AI), E.l.f. Beauty (cited) **Keywords** AI adoption, return on investment, process transformation, employee training, data, TCS study 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 Merca2 and its author. Entry translated from French by Claude Sonnet 5.5.