Why AI Isn’t Changing Core Business Yet

Most big firms have not turned AI into the heart of their work. Instead they stick it on the side, like using it for meeting notes or simple marketing tasks. Only a handful have tried to let AI reshape the main processes that make money.

Changing the core is risky, similar to early factories that first used electric power just for lighting before they rebuilt the whole line around motors. Experts say the real gain comes from rethinking the whole workflow, not just slapping AI onto old steps. At Mastercard, teams redesigned product development to be faster and smaller, while AXA moved from measuring many AI ideas to focusing on a few that could deliver steady value.

When an AI agent messes up, responsibility is shared. The business owner looks at value, the technologist checks function, and the security officer watches for risks. Leaders suggest not spending time on blame unless there is a serious breach; instead they encourage teams to try, learn from failure, and keep improving.

AI tools lose usefulness quickly. What was useful a year ago can feel outdated today, prompting constant rebuilds. AXA, for example, plans to earn between five hundred and seven hundred million euros each year from AI by 2029, while sitting on a pile of four billion unstructured documents that still need work.