
In January this year, Visa, the global card-payment company, announced that it would start using artificial intelligence (AI) to handle disputed card transactions. Instead of calling a customer-service agent, customers will be able to interact with a bot that, in most cases, will be able to resolve their case without human intervention.
The technology that Visa intends to use for this is known as “agentic AI”—AI that can make decisions and execute a series of tasks with little or no human intervention.
It is the third in a wave of AI innovations that have emerged over the past decade or so, following on from predictive AI, which makes recommendations and predictions by crunching large amounts of data, and generative AI, such as ChatGPT, which can produce text responses to natural-language queries.

The relatively recent arrival of agentic AI is symptomatic of how fast AI innovation is taking place. Little more than a year ago, few people outside of AI labs had heard of it. Today the agentic-AI market is worth about $2.6bn, and it is forecast to grow to almost ten times that by 2030. And, says Richard Murphy, one of the authors of a new report on the use of AI in companies and public bodies, that speed of innovation means many organisations are finding it hard to keep up and deploy AI effectively at scale.
The report, “Enterprise Maturity Index 2025”, was jointly produced by ServiceNow, an enterprise cloud software company headquartered in the United States (where Mr Murphy is head of global thought leadership research), and Oxford Economics, a consultancy based in the UK. It surveyed directors, senior directors and the C-suites of 4,473 organisations in 16 countries to hear how their enterprises are progressing towards what is known as “AI maturity”: how well they are using AI and how much they are benefiting from it.

While 55% of organisations surveyed said they had rolled out more than 100 different AI-use cases, only 29% had a defined set of metrics to measure the return on that investment.

The Index defines AI maturity across five pillars: strategy and leadership; workflow integration; talent and workforce; governance; and realised value. Most enterprises falter on multiple fronts.
Talent is a widespread pain point. Only 33% of organisations strongly agreed that they had the right mix of skills to deliver on their AI strategy. While training initiatives are expanding, most companies are still in the process of identifying the skills they need.
Governance is equally fraught. Employees are being encouraged to experiment with AI, but only 44% of firms have a designated team that drafts AI policies, mitigates AI risks and focuses on responsible AI use. According to Dorit Zilbershot, ServiceNow’s group vice-president, AI experiences and innovation, and another of the report’s authors, this will only become more of an issue as AI grows in complexity. “The autonomous capabilities of agentic AI only increase the need for robust guardrails,” she explains.
The report also found that many organisations are unclear about how effectively their AI budget is being spent. While 55% of organisations surveyed said they had rolled out more than 100 different AI-use cases, only 29% had a defined set of metrics to measure the return on that investment.
The research does highlight, however, a small group of “pacesetter” organisations that are outperforming the rest of the group and from whom, Mr Murphy says, everyone can learn lessons.
Pacesetter organisations made up just under 20% of the organisations surveyed but their average AI maturity score was 44, compared with an average of 35 for the cohort as a whole. And they reported significantly higher returns from their AI strategies. Over 80% of pacesetters reported an increase in gross margins, compared with 64% of the other organisations surveyed. More than half were seeing increased efficiency and productivity and the ability to innovate faster, compared with under a third of the non-pacesetter group.

Organisations need to move away from a legacy mindset of merely iterating on yesterday’s technologies to an AI-first one.
Brian Solis, head of global innovation, ServiceNow

These high-performing organisations distinguish themselves in five areas. They proactively encourage AI innovation, often by creating a dedicated AI innovation centre. They use a single data platform across the organisation, which allows data to flow between business functions and eases the introduction of new technologies. They invest heavily in training and recruitment to foster AI talent. They take a proactive approach to data governance, which allows employees to experiment safely with potential AI solutions. And they show a readiness to adopt new technologies, such as agentic AI.

Brian Solis, head of global innovation at ServiceNow, acknowledges that these can feel like big steps for other organisations to take. But he believes they have the power to unleash vast hidden potential.
“Organisations need to move away from a legacy mindset of merely iterating on yesterday’s technologies to an AI-first one,” he says. “Such a mindset allows leaders to dig deeply into the structure of their organisations and see, sometimes for the first time, the root causes of the obstacles they face and how to solve them.”
All of this, of course, may be alarming to employees concerned about losing their jobs to an AI bot. But Ms Zilbershot is optimistic about a future in which AI and humans will be working in tandem.
“Instead of the traditional ‘human-in-the-loop’ approach, where humans are deeply involved in every step, we shift to a ‘human-on-the-loop’ model,” she says. “Employees will still need to play a proactive role in reviewing and validating outputs, as well as scrutinising the underlying processes and data sources that inform these outputs.”
Mr Solis agrees. He says that becoming an AI-powered organisation is about more than just deploying technology, good governance and appropriate metrics. It’s a matter of culture too. And here, again, the pacesetters are leading the way.
“In their approach to change management, they understand that trust, autonomy and agility are paramount to the future of work and business,” he says. “Even as they double down on machines, these organisations are keeping humans at the heart of everything they do.”
