AI adoption is moving faster than governance structures can manage at 95% of retail organisations, creating a growing divide between those making technology decisions and those ultimately responsible for the risks, according to new research.
Thoughtworks’ global survey of 3,200 CIOs found the governance gap is particularly pronounced in retail. While 62% of retail CIOs feel prepared to govern AI, 95% believe the pace of adoption is outstripping existing governance models.
Responsibility for AI strategy is also shifting beyond traditional technology teams. Some 62% of retail CIOs said CEOs are increasingly influencing AI decision-making, compared with 23% of organisations across the wider survey.
By contrast, just 7% of retail respondents identified central IT or technology leadership as having the greatest influence over AI decisions.
Despite this increasingly distributed decision-making, accountability remains firmly with technology teams when problems arise. Some 96% of retail CIOs said central IT remains responsible for AI-related security or compliance issues caused by tools purchased elsewhere in the organisation.
For retailers deploying AI across areas such as ecommerce, customer experience and wider business operations, the findings highlight the potential risks of allowing adoption to develop without equally clear decision-making and governance structures.
The wider survey found similar concerns around accountability. Some 90% of CIOs said they are held responsible for AI failures involving independently integrated tools, while 37% felt personally accountable for AI-related security incidents and 35% for data privacy breaches.
AI is also changing the role of technology leaders. Globally, 89% of CIOs said they are now more responsible for redesigning workforce workflows and labour models than managing core IT infrastructure, rising to 92% among UK respondents.
Mike Sutcliff, CEO of Thoughtworks, said businesses do not necessarily need to centralise every AI decision within IT, but require clearer decision rights and appropriate skills and information across their organisations.
For retail leaders, the research suggests that scaling AI successfully will increasingly depend not simply on adopting new tools, but on establishing who can approve their use, how risks are assessed and where accountability sits when things go wrong.
