Skip to content
Marsanne Studio
← The AI watch

Data study ·

Economic returns from AI remain low for businesses

Original title, in Spanish: “Rendimientos de negocio aún deficientes de la IA”

Publication
HayCanal (Spain)
Author
Unsigned
Published on
10 September 2026 · read in the page metadata

Summary

The Unisys AI & Cloud Insights Report 2026 shows a growing gap between organisations' confidence in their technology capabilities and the actual economic results of AI. According to the study, 90% of organisations say they have the right architecture for data-driven decision-making, up from 72% in 2025, but only 65% believe their operational efficiency exceeds expectations, down from 80% the previous year. Adoption of agentic AI is progressing more slowly than the enthusiasm it generates: 75% consider it essential for managing the cloud, but only 23% have deployed it across the organisation. Cybersecurity breaches have almost tripled, affecting 49% of organisations compared with 17% a year earlier. The main obstacles identified are the talent shortage (27%) and data management issues (26%). Mike Thomson, CEO of Unisys, calls for demonstrating a real return on investment.

Summary based on the full text of the article.

Read the article on HayCanal →

Figures cited

  • 90 %

    Share of organisations with an architecture suited to data-driven decision-making at scale (2026 vs 2025)

    Global organisations surveyed by Unisys, 2026 report (vs 72% in 2025)

    « el 90 % de las organizaciones afirma disponer de la arquitectura adecuada para respaldar la toma de decisiones basada en datos a gran escala, frente al 72 % registrado en 2025. »

    Source: Unisys AI & Cloud Insights Report 2026

  • 49 %

    Share of organisations that suffered a cybersecurity breach in the past twelve months

    Global organisations surveyed by Unisys, 2026 report (vs 17% the previous year)

    « casi la mitad (49 %) de las organizaciones sufrió una brecha de ciberseguridad en los últimos doce meses, una cifra que prácticamente triplica la del año anterior (17 %). »

    Source: Unisys AI & Cloud Insights Report 2026

  • 75 %

    Share of organisations considering agentic AI essential for managing cloud environments

    Global organisations surveyed by Unisys, 2026 report

    « aunque el 75 % considera que será esencial para gestionar los entornos cloud, sólo el 23 % ha comenzado a desplegarla a escala en toda la organización. »

    Source: Unisys AI & Cloud Insights Report 2026

  • 23 %

    Share of organisations that have started deploying agentic AI across the whole organisation

    Global organisations surveyed by Unisys, 2026 report

    « sólo el 23 % ha comenzado a desplegarla a escala en toda la organización. »

    Source: Unisys AI & Cloud Insights Report 2026

  • 96 %

    Share of organisations whose cloud security strategy improves their ability to adopt new technologies faster than competitors (2026 vs 2025)

    Global organisations surveyed by Unisys, 2026 report (vs 60% in 2025)

    « el 96 % afirma que su estrategia de seguridad en la nube mejora su capacidad para adoptar nuevas tecnologías más rápidamente que sus competidores, frente al 60 % registrado en 2025. »

    Source: Unisys AI & Cloud Insights Report 2026

  • 65 %

    Share of organisations whose operational efficiency exceeds expectations (2026 vs 2025)

    Global organisations surveyed by Unisys, 2026 report (vs 80% in 2025)

    « Solo el 65 % afirma que su eficiencia operativa supera las expectativas, frente al 80 % registrado en 2025. »

    Source: Unisys AI & Cloud Insights Report 2026

  • 27 %

    Share of organisations citing the talent shortage as the main obstacle

    Global organisations surveyed by Unisys, 2026 report

    « Superar desafíos como la escasez de talento (27 %) y los problemas relacionados con la gestión e integración de datos (26 %), las principales barreras identificadas por el estudio »

    Source: Unisys AI & Cloud Insights Report 2026

  • 26 %

    Share of organisations citing data management and integration issues as the main obstacle

    Global organisations surveyed by Unisys, 2026 report

    « los problemas relacionados con la gestión e integración de datos (26 %), las principales barreras identificadas por el estudio »

    Source: Unisys AI & Cloud Insights Report 2026

  • 42 %

    Share of executives who see the cloud and IT function as profit generators rather than mere cost centres

    Global organisations surveyed by Unisys, 2026 report

    « menos de la mitad de los directivos (42 %) considera que el cloud y la función de TI son centros de generación de beneficios, en lugar de simples centros de costes »

    Source: Unisys AI & Cloud Insights Report 2026

  • 77 %

    Share of organisations still at an early stage of agentic AI adoption

    Global organisations surveyed by Unisys, 2026 report

    « la mayoría (77 %) sigue encontrándose en fases iniciales de adopción »

    Source: Unisys AI & Cloud Insights Report 2026

  • 89 %

    Share of organisations that have identified priority use cases for agentic AI

    Global organisations surveyed by Unisys, 2026 report

    « el 89 % ya ha identificado casos de uso prioritarios; el 87 % está formando a sus empleados; otro 87 % ha conseguido el respaldo de la dirección, y el 86 % colabora con proveedores especializados en IA y tecnología. »

    Source: Unisys AI & Cloud Insights Report 2026

  • 93 %

    Share of organisations whose cloud security capabilities significantly influence the level of autonomy granted to AI systems

    Global organisations surveyed by Unisys, 2026 report

    « el 93 % afirma que las capacidades de seguridad en la nube influyen de forma significativa en el nivel de autonomía que conceden a sus sistemas de IA. »

    Source: Unisys AI & Cloud Insights Report 2026

Classification

Country
Worldwide scope
Company size
All sizes
Industries
All industries, or industry not specified
Companies
Unisys (study author)
Keywords
agentic AI, return on investment, cybersecurity, cloud, technology transformation, talent shortage

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 HayCanal. Entry translated from French by Claude Sonnet 5.5.