The numbers come first. At a Fortune Brainstorm AI virtual event, Marc Warner, Accenture’s global chief technology officer, said the number of large enterprises or government processes 'that have been really, truly reinvented with AI at the absolute core is basically close to zero.'
Warner said many companies have pushed AI into peripheral work such as marketing, meeting notes, and transcription. He said coding is 'perhaps a little different' because the returns can be measured. But he argued that changing the economics of an industry requires going after core business processes, which is risky.
He compared the current moment to early factories that used electricity only for lighting before redesigning production around electric motors. His point was blunt: adding AI around the edges is not the same as restructuring the business itself.
Warner was one of four executives on the panel. The others were Matthieu Caillat of AXA, Greg Ulrich of Mastercard, and Marjorie Janiewicz of Mistral AI. Their discussion centered on why companies report little return from AI despite broad use.
Ulrich said the biggest gains come from redesigning whole workflows, not layering AI on top of existing ones. He said Mastercard is 'very, very advanced' in redesigning its product development life cycle, with smaller teams and faster turnaround. He also warned that speeding up one step in an 80-step process can simply create a backlog at the next step.
'The AI agent to automate a piece of this isn’t ... the long pole in the tent for a lot of this,' Ulrich said. 'It’s the rethinking. It’s the remapping.'
Warner called AI transformations 'politics projects' and said he had only recently stopped thinking of them as technology projects. He still described the AI builds themselves as technology projects, but said the real barrier is organizational change.
Caillat made a similar point. As technology improves, he said, the work becomes less about software and more about 'How do you bring people along?' Ulrich said Mastercard leans on CEO sponsorship, including weekly updates with the CEO’s direct reports. He added that no one has solved the politics of large organizations.
Warner also said companies should not obsess over blame when AI goes wrong. He said there is no single clean answer on responsibility. In general, he said, the business owner is accountable for an agent’s value, the technologist for its functionality, and the chief information security officer for its security.
He urged firms not to dwell on the blame game yet, except in critical incidents such as a major security breach. Janiewicz agreed that finger-pointing is not useful. She said the better question is whether a company controls the technology it adopts and can set its own guardrails.
The market has already voted on the limits of easy AI adoption: the quick wins are mostly in support functions, not in the hard work of redesigning production. That is where productivity gains live, and where management failure can hide behind buzzwords.
Capital rewards clear rules, not politics projects. If AI is going to matter at scale, executives will have to tear down old workflows instead of decorating them with new tools. That is harder than buying software. It is also where the margin improvement, if any, will come from.
