The AI ROI Nobody Measures

IKEA's AI chatbot saved the company €13 million. That is not the number worth remembering. The number worth remembering is €1.3 billion, and it has almost nothing to do with cutting costs.

Nearly every conversation about AI adoption stops at the same place: hours saved. Faster workflows, fewer manual steps, time back on the calendar. That is real, and it is worth having. But it is the floor. What you do with those reclaimed hours is the ceiling, and most companies never touch it.

In 2021, Ingka Group (the largest IKEA franchisee) launched a customer-service chatbot called Billie. It went on to handle around 47% of incoming queries, freeing roughly 8,500 call-center employees from repetitive work. The obvious next move was to bank the savings, trim the team, and call it a win.

IKEA did something else. They looked at the half of questions the bot could not resolve and noticed a pattern: those customers did not want faster answers, they wanted design help. So instead of cutting 8,500 people, IKEA retrained them as remote interior-design advisors and built a paid advisory service around them.

That new service generated €1.3 billion in revenue in FY2022, about 3.3% of Ingka's total, with a target of reaching 10% by 2028. Same AI. Same people. Roughly one hundred times the return. The entire difference came from what they chose to do with the freed-up capacity.

Every AI pitch leads with hours saved. Every ROI deck is a cost-reduction story. So that is where most companies stop. They automate a workflow, pocket the time, maybe remove a role, and move on. The saved hours sit there as potential energy that never gets spent.

The companies winning with AI treat freed capacity as a budget: a pool of hours they can finally point at the growth work that never had bandwidth before. The launch that kept slipping. The customer segment nobody had time to chase. The new offer that never got finished.

McKinsey's research on AI high performers makes the same point. The companies seeing outsized returns are not the ones chasing efficiency hardest. They are the ones that set growth and innovation as explicit goals from the start. Efficiency-only strategies produce incremental gains; growth strategies move the actual business, and high performers report returns several times the average.

Accenture freed up roughly 57,000 hours a year in its finance team alone, then redeployed that capacity into growth initiatives and funding strategic moves. Not layoffs. Redeployment toward what creates value.

You do not need IKEA's scale for this to matter. If AI gives a six-person team back ten hours a week, that is not a cost saving. That is a part-time hire you did not have last quarter, pointed at whatever you have been too busy to build.

Which means the question most leaders ask about AI is the wrong one. "How many hours will this save me?" measures the floor. The better question is: "What would we finally build if those hours showed up on Monday?"

That is where the real return lives. Not in the time you get back, but in what you do with it. If you want help finding the workflows worth automating, and the growth work worth pointing the freed capacity at, that is exactly the conversation we have with the businesses we work with.