Companies question whether AI spending is paying off

In short: Businesses are spending heavily on AI, but many are still unsure how much value they are actually getting back.

What's going on

At the Financial Times Future of AI Summit, speakers described a growing gap between how much companies spend on AI and how clearly they can measure the results. “ROI” means return on investment, or what you get back compared with what you put in.

Alex Schultz, Meta’s chief data officer, said AI has made some parts of Meta more efficient, especially behind-the-scenes data work. But he suggested the benefits are not as clear in other parts of the company, even where AI is widely used. The message was that the same tool can help one team and slow down another.

The newsletter also pointed to Goldman Sachs as an example of this uneven impact. After reviewing internal productivity data, Goldman restricted AI coding tools to its engineers, and stopped bankers and traders from using them. The idea is similar to giving power tools to the people trained to use them, but not handing them out to everyone.

One reason the picture is blurry is how companies measure ROI. A Wharton business school report found 40 per cent of tech companies said they were seeing strongly positive ROI from AI, and 20 to 35 per cent in other sectors. But many firms measure success as “time saved,” which can ignore time spent checking AI output for mistakes.

What to watch

Speakers also discussed “AI agents,” meaning AI tools that can take actions for you, like making routine purchases (a bit like a digital assistant that does errands). That could change online shopping and customer service, but it may also create new friction if consumers’ agents and companies’ systems end up flooding each other with requests.

Source: Financial Times