AI adoption accelerates despite fears of an AI winter, says research

Agentic AI is gaining ground as leaders balance opportunity, risk and FOMO

New research released today by Capgemini indicates that enterprise leaders are looking beyond the hype surrounding AI and are taking a more pragmatic approach to its use. However, leaders remain shy about disclosing publicly its use in decision making.

The research, "The multi-year AI advantage: Building the enterprise of tomorrow" surveyed more than 1500 executives globally, 38% of organisations already operationalise generative AI use cases, while six in ten organisations are now exploring the application of agentic AI.

China is more advanced than the US and Europe with 46% of Chinese organisations are piloting or deploying agentic AI, as opposed to 40% in the US and 30% in Europe.

Researchers also found that data sovereignty is emerging as a critical priority for 2026 which is perhaps unsurprising given the sharp deterioration in relations between the US and Europe, and the tumultuous state of geopolitics more broadly.

54% prioritise data sovereignty, keen to ensure that their confidential data remains within their control, regardless of platform. AI sovereignty is subsequently gaining in importance as organisations understand the link between the data feeding these models and outcomes.

Nonetheless, the market for generative and agentic AI is still at least partly driven by FOMO. Two thirds of business leaders expressed a fear of being left behind strategically if their competitors got ahead in the race to scale up AI deployments.

The desire not to be left behind even overrode concerns about the prospect of an AI winter setting in. Most of those surveyed acknowledged the likelihood of the optimism and financial confidence in AI starting to ebb, but 80% said that they planned to wait it out by focusing on foundational capabilities such as the data and infrastructure. 72% planned to prioritise low cost, high impact initiatives. A smaller proportion (57%) said that they intended to keep hold of “core talent and knowledge.”

AI winter or not, organisations are planning to accelerate AI investments. They are just doing so more selectively, prioritizing functions with well-defined processes and measurable outcomes. On average, they expect to allocate 5% of their annual business budget to AI initiatives in 2026, up from 3% in 2025.

Capgemini’s researchers also note that organisations have evolved in the way that they measure AI success. It's not all about cost reduction. New measures of ROI include revenue growth, risk management and compliance, knowledge management and customer experience and personalisation.

AI is reshaping decision-making, but few will admit it

Additional research from Capgemini shines a light on executive decision making. "How AI is quietly reshaping executive decisions", surveyed 500 C-level executives, including 100 CEOs. It finds that more than half of this group uses AI to support or inform their strategic decision-making in some way.

Only 17% said that they “actively” use AI to support decision making. 42% did so “selectively”, and 31% are currently “experimenting” with it. However, the proportion expecting to use AI actively within one to three years is 38%.

Early adoption is already delivering value. More than half of executives reported reduced time and cost to make decisions, and improvements in creativity and foresight from the use of AI.

However, leaders are clear that AI should not be considered a replacement for human judgement – wisely given the propensity of generative AI to make mistakes or just make things up. Only 41% described their level of trust in AI for executive decision-making as “above-average.” Leaders were mindful of legal and security risks, and the difficulty of explaining decisions to others when they were AI-influenced.

A particularly interesting finding of this piece of research is a dichotomy between private enthusiasm for using AI, and public reticence to admit doing so. Only 11% of executives were prepared to highlight publicly the use of AI in business decisions.

Why the shyness around such a transformative technology? Those who prefer not to disclose AI use cite concerns about reputational risk if AI-influenced decisions go wrong and uncertainty around how other people - clients, partners and the public - perceive AI use.