Field Note
CEO Peer Pressure and AI Adoption
When everyone does it, not following the crowd has consequences.
In January 2026, the Federal Reserve asked banks a question about AI that every CEO should have noticed.
It was not a philosophical question. It was not about AGI, labor replacement, or whether a model could reason. It was a lending question.
Compared with January 2025, were banks more or less likely to approve commercial and industrial loans to companies with different levels of exposure to artificial intelligence?
The question appeared in the Senior Loan Officer Opinion Survey on Bank Lending Practices, usually shortened to SLOOS. The name is dry enough to disappear into the wallpaper, which is unfortunate, because the survey is one of the cleaner places where institutional belief becomes visible.
The Fed uses SLOOS to ask senior loan officers whether lending standards are getting tighter or looser, whether demand for loans is rising or falling, and what kinds of borrowers are becoming easier or harder to finance. It is not a technology survey. It is a credit survey. That is why the AI question matters.
The January 2026 answer was direct enough to matter.
For firms in sectors that banks believed would benefit from high AI exposure, 22.2 percent of banks said they were somewhat more likely to approve a loan. Among large banks, the number was 42.1 percent. For firms in sectors that banks believed would be adversely affected by high AI exposure, 58.5 percent of banks said they were less likely to approve a loan. For firms with little AI exposure, every bank in the survey said its likelihood of approval was basically unchanged.
A credit officer does not need to believe in every AI claim for this to matter. The question itself means AI had moved from pitch deck to credit file.
A CEO does not have to believe every AI claim to understand what just happened. She does not have to believe the consultant. She does not have to believe the board member who came back from a conference suddenly convinced the entire company needs an AI transformation office. She does not have to believe the investor who talks about productivity gains as if execution risk has been repealed.
But she does have to believe her lender.
Once banks begin asking whether AI exposure changes loan approval, AI becomes part of the capital-access story. It may not be formal in every credit model. It may not be weighted the same way by every lender. It may not even be right in every industry. The point is narrower. The question has entered the room.
When that happens, silence starts to carry meaning.
If the CEO has no AI strategy, the bank may hear complacency. If the CEO has no pilots, the board may hear drift. If the CEO says the technology is overhyped, investors may hear that management is under-reading the next productivity shift. If the CEO waits too long, employees may hear that the company is going to be late to the tools that reshape their own work.
That is the peer pressure in this piece. Not the high school version, where the consequence is embarrassment. This is the adult version, where the consequence shows up in credit terms, valuation, analyst patience, board confidence, hiring, and competitive narrative. The room stays polite. The pressure still works.
The CEO may privately think the AI market is overheated. She may be right. She may think half the vendor claims are old workflow automation in new clothes. She may also be right. She may know that many AI pilots will fail, that bad implementations will create security and compliance problems, and that the wrong architecture can leave a company with cost, complexity, and no operating gain worth naming.
None of that removes the pressure to gesture. So companies start pilots. They form AI councils. They hire advisors. They ask every function to identify use cases. They run internal copilots. They brief the board. They add AI sections to investor presentations. They tell employees they are experimenting responsibly. They tell lenders they are adapting. They tell customers they are not asleep.
Some of this will become real work. Some of it will become theater. Most of it will live in the awkward middle, where organizations usually live when the future has not resolved but the social cost of inaction has already arrived.
The rational move, in that middle, is not blind transformation. It is visible learning. Do enough that capital providers, boards, and employees can see motion. Avoid doing so much that the company locks itself into the first architecture, vendor stack, or management theory that happens to arrive with a polished deck and a procurement form.
The old advice was not to be first and not to be last. That sounds cowardly until you sit in the CEO seat. Being first into the wrong answer can waste capital, burn credibility, and lock the company into a bad architecture before the market has settled. Being last to the right question can make management look asleep.
That is the CEO's problem now.
The old question was whether AI worked well enough to justify adoption. The new question is whether a company can afford to look like it has no theory of adoption at all.
The SLOOS question did not settle the AI debate. It did something more interesting. It showed that banks had begun sorting borrowers through an AI lens. That lens may be crude. It may over-reward some companies and over-penalize others. It may change as evidence arrives.
But it exists.
And once a lender asks the question, the CEO can no longer pretend the question belongs only to the technology team.
Sources
- Federal Reserve, January 2026 Senior Loan Officer Opinion Survey summary.
- Federal Reserve, January 2026 SLOOS Table 1, questions 41 and 42.