Agentic AI Goes Mainstream as Majority of Companies Report Leveraging Multiple AI Agents
London, UK: 14th October 2025 – Leaders are adopting AI so widely that they now trust it to take charge of critical business decisions, with three-quarters (76%) of UK executives believing AI has become so essential their company would struggle to function without it.
This is according to new survey data from PagerDuty, a global leader in digital operations management. The company’s latest AI Resilience Survey analyses insights from 1,500 IT and business executives in the UK, US, Australia, Japan, France and Germany.
In the UK, a vast majority of executives (88%) trust AI agents to take action on behalf of their organisation during a crisis, such as an outage or security event. The UK trust level is higher than the international average (81%).
Growing Maturity of AI Deployment
When it comes to the fast-growing area of agentic AI and AI agents, executives on both the business and IT side are increasingly confident that AI can protect their interests, even when things go wrong.
In just one year, acceptance of AI has increased widely among business and IT executives. In the UK, more than three out of four (78%) trust AI-generated outputs more than they did a year ago. The rising trust sentiment is highest in France (81%), but not as strong in Japan, where trust is lower (69%).
More than half of UK executives (54%) cite a rise in the quality of AI outputs while almost as many (52%) are using AI more frequently with positive results. More than half of UK executives (52%) have gained a better understanding of how AI works, while the same amount of UK executives (52%) are looking at successful AI use in other teams or at other companies.
AI as Catalyst for Coding Efficiency
One of the reasons AI is seen as so crucial by companies is that an increasing amount of code is being created or reviewed with the help of AI. More than four out of five (85%) UK organisations are using AI to write, review or suggest code.
AI coding is on the rise, but companies still feel it’s necessary to make sure that technology is doing the job well. To that end, they’re developing processes to test AI-generated code for accuracy and reliability. A vast majority of UK organisations (88%) are doing some form of AI code testing, but less than half (47%) do it consistently and through a formal process. The US leads in formal and consistent testing at a significantly higher 59% while Japan lags at 19%.
AI Confidence Tempered by Caution
Despite growing AI reliance and trust, executives are also clear-eyed enough to see that AI isn’t perfect and that there are reasons to have a few doubts about AI models. Nearly two out of five UK executives (38%) cite news they’ve read about AI models or regulations that have made them trust-AI generated outputs less in the last year.
Additionally, more than a quarter of UK executives (27%) worry about a lack of transparency in how AI systems generate outputs, and 35% are concerned about low-cost, lower-quality AI models becoming more widely available.
Eduardo Crespo, VP EMEA, PagerDuty said: “AI has shifted from sidekick to decision-maker, with AI agents being trusted to take action during the most business-critical moments. What’s striking is that UK leaders, in particular, are placing a higher-than-average level of trust in AI agents to take action during outages or security incidents,which is a sign of how quickly organisations are moving from AI experimentation to operational reliance. But trust doesn’t mean blind faith. The businesses leading in AI adoption are the ones balancing confidence with caution, building governance around their AI use while leaning into the speed, scale and resilience it can unlock.”
Methodology
The PagerDuty AI Resilience Survey was conducted by Wakefield Research among 1,500 IT and Business Executives, with a minimum seniority of Director, at companies with a minimum annual revenue of $500m USD. The research was conducted in the following six markets: Australia, France, Germany, Japan, U.K. and U.S., with quota set for 250 respondents per market, within each market a 50/50 quota was set for IT Executives and Business Executives, between July 7th and July 15th, 2025, for the U.S., U.K., Australia and Japan, and July 31st to August 7th, for France and Germany, using an email invitation and an online survey.
Results of any sample are subject to sampling variations. The magnitude of the variation is measurable and is affected by the number of interviews and the level of the percentages expressing the results. For the interviews conducted in this particular study, the chances are 95 in 100 that a survey result does not vary, plus or minus, by more than 2.53 percentage points in the global sample, and 6.2 percentage points in each market from the result that would be obtained if interviews have been conducted with all persons in the universe represented by this sample.
