--- title: "AI and tech myths – Debunk #5: \"Enterprise AI means generative AI\" — Marsanne Studio AI watch" description: "An opinion piece argues, citing figures from McKinsey, IDC, Gartner and BCG, that tabular AI carries more weight than generative AI in business." url: "https://marsanne.studio/en/ai-watch/ai-and-tech-myths-debunk-5-enterprise-ai-means-generative-d44ebd/" 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) Opinion · 8 October 2026 # AI and tech myths – Debunk #5: "Enterprise AI means generative AI" Original title, in French: “Les mythes de l’IA et de la tech – Debunk #5 : « L'IA en entreprise, c'est l'IA générative »” **Publication** Maddyness (France) **Author** Arnaud Lusetti **Published on** 8 October 2026 · read in the page structured data ## Summary In this opinion piece, Yann Lechelle, co-founder of Probabl, argues that tabular AI, which makes decisions from data in rows and columns (credit, fraud, pricing, forecasting), generates most of the value of AI in business, far more than generative AI. He draws on figures from McKinsey (generative AI potential estimated at $2,600-4,400 billion a year versus $11,000-17,700 billion for non-generative AI; 88% of companies use AI but only 39% see an impact on earnings), IDC (17% of global AI spending on generative AI, with a projected 32% in 2028), Gartner (90% of "airtime" versus 5% of real use cases) and BCG (62 to 70% of value in core functions). The author also discusses recent advances (TabPFN, CARTE, TabICL, SAP's acquisition of Prior Labs) and calls on executives to rebalance their AI investments. Summary based on the full text of the article. [Read the article on Maddyness →](https://www.maddyness.com/2026/10/08/mythes-ia-debunk-5-ia-tabulaire/) ## Figures cited - — % Share of unstructured enterprise data according to IDC Companies, IDC estimate > « IDC estime que 80 à 90 % des données d'entreprise sont non structurées » Source: IDC - 60 % Share of technology spending that supports structured data Corporate technology spending, IDC > « Le même institut relève qu'environ 60 % des dépenses technologiques soutiennent déjà la minorité structurée » Source: IDC - — USD billion Annual potential of generative AI estimated by McKinsey in 2023 Worldwide, 2023 > « le chiffre qui a lancé la ruée date de 2023 : McKinsey évaluait le potentiel de l'IA générative entre 2 600 et 4 400 milliards de dollars par an » Source: McKinsey - — USD billion Annual potential of non-generative AI and analytics according to McKinsey Worldwide > « ce montant s'ajoute aux 11 000 à 17 700 milliards attribués à l'IA non générative et à l'analytique » Source: McKinsey - 17 % Share of generative AI in global AI spending Global AI spending, IDC > « IDC situe l'IA générative à 17 % des dépenses mondiales en IA, avec une projection à 32 % en 2028 » Source: IDC - 32 % Projected share of generative AI in global AI spending in 2028 Global AI spending, 2028 projection, IDC > « IDC situe l'IA générative à 17 % des dépenses mondiales en IA, avec une projection à 32 % en 2028 » Source: IDC - 90 % Share of "airtime" taken up by generative AI according to Gartner Gartner analysis > « un analyste de Gartner a livré la formule la plus juste : l'IA générative, c'est 90 % du temps d'antenne et 5 % des cas d'usage » Source: Gartner - 5 % Share of enterprise AI use cases that involve generative AI according to Gartner Gartner analysis > « un analyste de Gartner a livré la formule la plus juste : l'IA générative, c'est 90 % du temps d'antenne et 5 % des cas d'usage » Source: Gartner - — % Share of data professionals working primarily with relational data Kaggle survey of data professionals > « dans l'enquête de référence de Kaggle, deux tiers des professionnels de la donnée travaillent d'abord sur des données relationnelles, et plus de 80 % dans la distribution et l'assurance » Source: Kaggle - — % Share of AI value located in core business functions according to BCG Core functions: operations, pricing, supply chain, R&D > « le Boston Consulting Group situe entre 62 et 70 % de la valeur de l'IA dans les fonctions cœur : opérations, tarification, chaîne d'approvisionnement, R&D » Source: Boston Consulting Group - 88 % Share of companies reporting that they use AI Companies, McKinsey state of AI survey > « 88 % des entreprises déclarent utiliser l'IA, mais 39 % seulement constatent un effet sur leur résultat opérationnel, selon le dernier état des lieux de McKinsey » Source: McKinsey - 39 % Share of companies seeing an impact of AI on their operating profit Companies, McKinsey state of AI survey > « 88 % des entreprises déclarent utiliser l'IA, mais 39 % seulement constatent un effet sur leur résultat opérationnel, selon le dernier état des lieux de McKinsey » Source: McKinsey - 255 USD million Amount raised by Fundamental Fundraising by Fundamental (United States) > « tandis que Fundamental levait 255 millions de dollars » - 400,000 tables Number of tables in an ERP according to SAP Typical ERP, example cited by SAP's CTO > « Il évoquait les quelque 400 000 tables d'un ERP, en finance, chaîne d'approvisionnement, logistique et achats » ## Classification **Country** France, Germany, United States **Company size** All sizes **Industries** Banking, insurance, asset management, finance departments · Retail and distribution, e-commerce · Industry and manufacturing · Software vendors, IT services companies, telecoms · All industries, or industry not specified **Companies** Gartner (study author), Inria (cited), Fundamental (supplier), McKinsey (study author), SAP (cited), Prior Labs (supplier), Boston Consulting Group (study author), Kaggle (study author), IDC (study author), Probabl (supplier), Google (supplier) **Keywords** tabular AI, generative AI, structured data, credit scoring, fraud detection, tabular foundation 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 Maddyness and its author. Entry translated from French by Claude Sonnet 5.5.