Bain & Company just put a number on something most business leaders already feel in their gut: AI is not just another software upgrade, it is reshuffling who makes money and who does not. The consulting firm's new report says AI will move $4.7 trillion in global business profits between now and 2035.
To understand why that number is big, compare it to the last time a technology reshuffled the economy this way. The internet shifted about $1.4 trillion in profits over 20 years, mostly by making it cheaper for companies to reach customers and distribute goods. Bain expects AI to move more than three times that amount of money, in half the time.
The reason is the internet mostly changed how businesses sold things. AI changes how things get made. That means it reaches into places the internet never touched, like factory floors, hospitals, and drug labs, not just marketing departments and call centers.
Bain also found that 71 percent of business sectors will go through real structural change because of AI, compared to 41 percent during the internet era. If you run a company in almost any sector, from steel to insurance to logistics, this is not a trend you can sit out.
Here is the part that should change how leaders think about their AI budget. Bain splits the $4.7 trillion into two buckets. Straightforward productivity gains, like using AI to write emails faster or summarize documents, account for $1.1 trillion. The other $3.5 trillion, three quarters of the total, comes from innovation and shifts in market share, meaning companies that use AI to build new products, enter new markets, or steal customers from slower rivals.
Chasing generic productivity tools is the easy, visible move. But Bain's numbers say the real money is in figuring out what your company can now do that it could not do before, not in doing the same old tasks slightly faster.
There is a catch worth taking seriously before anyone gets too excited about productivity numbers. Glean's Work AI Institute surveyed thousands of office workers and found automation saves people about 11 hours a week on average. But much of that saved time gets eaten up by a new kind of work: checking AI's answers, fixing its mistakes, and feeding it the right information to begin with. Researchers gave this a name, "botsitting," and found workers spend close to 6 and a half hours a week doing it. Some workers, when they get tired of babysitting the AI, start shipping its unverified output anyway, a habit researchers call "botshitting."
Put together, these two reports tell a clear story. The profit opportunity from AI is real and enormous, but it will not show up automatically just because employees have AI tools open on their screens. It shows up when a company redesigns how work actually gets done, and points AI at specific goals tied to winning market share or building something new, not just at making existing tasks marginally faster.
Bain named several sectors where nobody has a locked-in advantage yet, including insurance, manufacturing, advertising, consulting, corporate law, and logistics. That is the opening. Whoever in these industries figures out the real use of AI first, rather than just adding more AI seats, will likely be the one who takes profit share from everyone else.