Companies keep pointing AI at their payroll. The real opportunity is pointing it at customers.

Every earnings call this year has some version of the same line: AI will let us do more with fewer people. I'm job hunting in 2026, so I hear it as a threat. As a PM, I hear it as a failure of imagination. So far, AI's biggest change has been to what consumers do all day.

Look at where people actually use it. An NBER study by OpenAI researchers and academic economists tracked ChatGPT from launch through July 2025 and found non-work messages grew from 53% to more than 70% of all usage. Pew found 49% of US adults now use AI chatbots, up from 33% in 2024, and 66% of adults under 30. Most of the action isn't at a desk. It's people using AI in their own lives.

Now the part that should excite every product person. Bank of America's payments data shows only about 3% of its households pay for AI services, though the number of paying households is up 38% from the 2024 average. About half of adults use it. Almost nobody pays for it yet. That isn't a saturated market. It's the biggest unbuilt consumer product category in years.

Meanwhile, the companies that went hardest at replacing people are walking it back. Klarna once said its AI assistant did the work of 700 support agents. In May 2025 its CEO told Bloomberg that "investing in the quality of human support is the way of the future," and that customers should always be able to reach a human. Commonwealth Bank of Australia cut 45 call center roles for an AI voice bot, then reversed the cuts in August 2025, admitting the roles "were not redundant" and apologizing to the staff.

The research points the same way. Stanford's Digital Economy Lab found no evidence of widespread, economy-wide job displacement from AI. Where it did find damage, it was specific: jobs for young workers fell in occupations where AI substitutes for human tasks. Where AI complements workers, employment was flat or rising. Replacement shrinks a team. Augmentation keeps it.

I learned what actually moves a business building consumer fintech. The feature that lifted on-time repayment about 11% in a mobile lending app wasn't a cost cut. It was pre-due-date reminders and one-tap repayment, designed from how borrowers actually behaved. Customer value compounds. Headcount savings happen once.

So if you're a PM with an AI budget, point it outward. Ask what your customers are already doing in a chatbot that your product should do better. Build for the life outside work, where most of the usage already is. Then make it worth paying for, because that 3% is the whole opportunity.

I built an AI job-search agent for myself this year. It didn't replace anyone. It cut my research time per application about 60% and let me spend that time on the parts only I could do. That's the version of AI I want to build at scale.

The easiest use of AI is cutting people. The valuable one is building something people will pay for.