Four in 10 small business employers have turned down a potential employee in the past year because their AI skills fell short. Among Gen Z respondents, it’s nearly two in three.
That comes from a new Patriot Software survey of 1,000 U.S. small business owners and hiring decision-makers. It inverts the usual AI jobs story, wherein the national conversation asks whether AI will take jobs. Small business employers already use it to decide who gets one.
The timing isn’t random. Demand for AI skills in entry-level jobs has nearly tripled since fall 2025, according to the National Association of Colleges and Employers. The class of 2026 is now entering the job market.
- 40.4% of employers have rejected a candidate in the past year over weak AI skills. Among Gen Z employers, it’s 65.3%.
- 17.7% now treat AI skills as a hard hiring requirement. Another 40.2% say the skills give a candidate an edge.
- 56.4% say AI-generated resumes have made it harder to spot qualified candidates.
- 44.4% offer no AI training of any kind. The top barrier is that it’s “not a priority” (24.6%), while cost ranks last (7.3%).
- 63.7% of small businesses use ChatGPT and 58.5% use Google Gemini as their AI tools.
- 66% would pay a premium of 5% or more for proven AI skills.
- 71.3% believe AI is splitting the workforce into workers who adapt and workers who fall behind.
- AI-use skews male among these employers: 42.7% of men use it multiple times a day, versus 29.9% of women.
Small business owners value AI skills in the hiring process
Four in 10 employers have rejected a candidate over AI skills. The rejections aren’t one-offs. Of the 40.4% who passed on a candidate over AI skills in the past year, about half did it more than once.
Younger employers enforce the bar hardest, which flips the usual assumption. 65.3% of Gen Z respondents have rejected a candidate over AI skills, against 47.3% of Millennials and 26.3% of Gen X.
The survey shows the pattern without pinning down the reason. Younger employers may run more digital businesses, or they may expect the fluency they take for granted in themselves.
Behind the rejections is a bar that moved: 17.7% of employers now treat AI skills as a hard requirement, and another 40.2% say they give a candidate an edge. Only 14.7% call AI irrelevant to the roles they hire for.
The skills in question aren’t exotic. These are the everyday tools employers already use themselves. 63.7% use ChatGPT and 58.5% use Google Gemini. Another 23.5% use AI built into their accounting or payroll software. Only 10.6% use none at all.
The AI shift reaches past small firms. The National Association of Colleges and Employers found that more than a third of entry-level jobs now require AI skills.
Employers demand AI skills, but most won’t train for them
AI skills are in demand. At the same time, employers don’t want to be fooled by AI reliance. 56.4% of employers say AI-generated resumes and cover letters have made it harder to identify qualified candidates.
By 2025, job seekers were submitting roughly 11,000 applications a minute on LinkedIn, up about 45% in a year, according to reporting by The New York Times.
When every resume reads as polished, employers stop trusting the paper. Many now test skills directly instead of taking claims at face value. That’s part of why the bar keeps rising. If you can’t trust the resume, you look for proof.
Employers will pay more for proven AI skills
AI skills are hard to find and easy to fake. The skills a candidate can prove stand out.
Two-thirds of employers (66%) say they’d pay a premium of 5% or more for a candidate who can prove strong AI skills. About 1 in 11 would pay 20% or more. Only a quarter would pay nothing extra.
The market backs this up. Lightcast, analyzing 1.3 billion job postings, found that listings calling for AI skills advertised salaries about 28% higher, close to $18,000 more a year.
More productivity doesn’t always come with more pay
Employers will pay to hire the skill. Rewarding it once it’s on the payroll is a different question.
When an employee masters AI and increases productivity, 28.5% of employers say that worker gets rewarded with better pay or a promotion. Another 29.8% hand that worker more responsibility.
The reward isn’t guaranteed. 26.6% say the more productive worker is simply expected to do more work. Another 16.4% say the workload grows with no matching pay.
Getting faster can mean a raise. It can also just mean a fuller plate.
Employers lean on AI heavily, but not evenly
Employers use these tools themselves, and heavily. 72.3% use AI at least a few times a week to run or grow the business, and 36.3% use it multiple times a day.
The habit isn’t even. 42.7% of men use AI multiple times a day, compared with 29.9% of women. And 14.0% of women say they never use it, more than double the 6.2% of men who say the same.
That’s a gap among the decision-makers themselves, before it ever reaches the people they hire.
An AI divide is forming in small businesses
Most employers can already see a divide forming. 71.3% agree that AI is splitting the workforce into workers who adapt and workers who fall behind. About a quarter (24.8%) say they already see it in their own business.
Employers also know the worker they want. Asked who will matter most over the next three years, 33.3% named a tech-forward generalist who works alongside AI fluently. That was the top answer, well ahead of the deep specialist at 13.8%.
Small employers expect a two-tier workforce and demand the skill that sorts people into it. Most have decided not to teach it.
Methodology
Patriot Software surveyed 1,000 U.S. small business owners and hiring decision-makers in July 2026 through Pollfish. Participants answered questions about how they evaluate AI skills, screen candidates, train employees, and reward productivity. Responses were analyzed by role, gender, generation, household income, education, and ethnicity.
Fair use policy
Users are welcome to utilize the insights and findings from this study for noncommercial purposes, such as academic research, educational presentations, and personal reference. When referencing or citing this article, please ensure proper attribution to maintain the integrity of the research. Direct linking to this article is permissible, and access to the original source of information is encouraged.
For commercial use or publication purposes, including but not limited to media outlets, websites, and promotional materials, please contact the authors for permission and licensing details. We appreciate your respect for intellectual property rights and adherence to ethical citation practices. Thank you for your interest in our research.
This is not intended as legal advice; for more information, please click here.


