Executives say AI returns depend on rebuilding business processes

Executives at Fortune Brainstorm Tech in Aspen said companies are struggling to get returns from artificial intelligence because many are adding AI to flawed systems rather than rebuilding how work gets done. The message matters for businesses spending heavily on AI tools while investors and managers ask when the technology will show up in revenue, profit and productivity.

At a Fortune Eye on AI breakfast, Manoj Bohra, chief technology officer at State Street, said companies first need what he described as foundation work, especially in regulated industries. According to Fortune, Bohra pointed to cleanly organized data, governance controls and detailed mapping of workflows as prerequisites before automation can produce reliable results.

Bohra said companies should not expect that groundwork to pay off within one or two years. He compared the investment to infrastructure projects whose value is judged over a longer period, according to Fortune.

Executives point to process design

Bill Briggs, Deloitte’s chief technology officer, said many companies have raced to roll out AI at scale without deciding what strategic goal the technology should serve. Fortune reported that Briggs said some AI projects do little for companywide revenue or profit because they target the wrong work.

Briggs also said companies often insert AI into existing processes rather than redesigning those processes around the technology. In his view, that can amplify inefficiencies, echoing the early period of industrial electrification when companies saw limited productivity gains after replacing steam power with electric equipment without changing factory design.

Kathy Pham, head of AI at ReviveHealth, also urged companies to start with the purpose behind a process. Fortune reported that Pham used bedtime stories as an example: AI might help if the goal is getting a child to sleep, but it would miss the point if the goal is focused time between parent and child.

Pham said business processes can drift away from their original purpose over time. Adding AI to those processes may fail to create value if companies do not first decide what outcome they want.

Some see near-term limits and selective wins

Stephen Balaban, cofounder and chief technology officer of AI infrastructure company Lambda, said AI is not ready for many corporate uses outside software development, according to Fortune. He said companies may be making a mistake by pushing AI agents broadly across large organizations today, while also arguing they should prepare for models to become capable enough for more domains within the next year or two.

Balaban also said businesses are right to press services firms to charge based on outcomes rather than staffing levels. That shift would tie AI work more directly to measurable business results.

Wen Sang, cofounder and chief operating officer of Genspark, said companies should target easy wins that increase revenue. Fortune reported that Sang cited advertising firms using AI to make video prototypes for pitches instead of paying artists to produce static storyboards, lowering costs while improving the chance of winning new work.

Faraz Shafiq, Wells Fargo’s chief AI product officer, said the bank has built horizontal AI components across business lines, including a unified agent platform and governance infrastructure. Within each line of business, he said, Wells Fargo works with domain experts to redesign processes from end to end.

Shafiq said some returns are straightforward to measure, citing a 25% increase in new account openings tied to AI tools. Other benefits, such as bankers spending more time with customers and building stronger relationships, may take years to convert into revenue, according to Fortune.

This story draws on original reporting from Fortune.