A July survey from Careerminds, a workforce consulting company specializing in career transitions and employee development, found that 67% of surveyed Arizona small-business owners believe artificial intelligence could perform a meaningful portion of the work done by at least one employee. The company projected that roughly 937,000 Arizona workers are in roles where AI could already perform part of the job.
The finding points to real labor-saving potential, but it should not be read as a forecast of immediate layoffs. Careerminds says the estimate is based on survey responses combined with employment data and small-business workforce shares. It measures perceived capability, not the number of positions expected to disappear. Separately, cyber researchers have warned that AI-driven attacks are creating new risks for small businesses, a development not measured in the Careerminds survey but relevant to any owner calculating the full cost of AI adoption.
Careerminds survey points to AI potential, not planned layoffs
Careerminds surveyed 3,022 small-business owners across the United States in July. The company said it used demographic quotas and weighting intended to make the sample broadly representative by age, gender, region and business size. However, the company did not disclose how many respondents were from Arizona or provide a state-level margin of error for the Arizona-specific figures.
That limits how far the state-level result can be taken. The Arizona figure supports a cautious conclusion: many small-business owners believe AI could take over meaningful pieces of existing work. It does not establish that those owners plan to dismiss employees, that the projection applies evenly across industries, or that every role counted by the estimate is at equal risk.
The same distinction matters for broader small business AI adoption decisions. A tool that can perform a task is not the same as a tool that can replace a job, and a job that can be partially automated may still require human judgment, customer relationships and company-specific knowledge.
Worker loyalty turns automation into a human decision
Among Arizona respondents, one-third cited loyalty as the main reason they would not replace an employee with AI. Seventeen percent pointed to the employee’s length of service, 13% cited company knowledge that would be difficult to replace, 12% said they felt responsible for an employee’s financial well-being, and another 12% said customers valued their relationship with the worker.
Nationally, 61% of respondents said they had delayed introducing AI because of concerns about its effects on employees, while 59% said they had rejected a specific AI product because its adoption could lead to job losses. At the same time, 56% expected AI to reduce the number of people their businesses employ within three years. Careerminds noted that some reductions may happen through attrition, as owners choose not to replace workers who retire or resign, rather than through immediate layoffs.
Amanda Augustine, a career coach and career expert for Careerminds, said personnel decisions are especially complicated for small-business owners because they often work directly beside employees and know their families and financial circumstances. She said owners know the people behind every paycheck and understand what job loss could mean for someone who has worked with them for years.
Replacing people with AI can weaken trust, not just payroll
The human impact of replacing employees with AI extends beyond the immediate loss of income for one worker. In a small business, a long-serving employee may carry informal knowledge about customer preferences, supplier habits, seasonal demand, internal workarounds and the owner’s expectations. That knowledge rarely appears in a software cost comparison, but it often affects how smoothly a business operates.
Customer relationships are another nonfinancial cost. A neighborhood business may rely on familiar staff to reassure repeat customers, solve unusual problems and recognize when a request needs personal attention. If that work is shifted to an AI tool without a human fallback, the business may save time while reducing the trust that made customers return in the first place.
There is also a community effect. Small businesses often provide local jobs where employees build skills, stability and professional identity. Replacing those roles with software can move savings onto the owner’s balance sheet while pushing the social cost onto workers, families and local labor markets. The Careerminds survey does not measure those downstream effects, but its loyalty and financial-well-being responses show that many owners are already thinking beyond subscription pricing.
Cyber exposure complicates the savings case
The Careerminds survey focuses on workforce tradeoffs, not cybersecurity. Still, AI deployment can create new risks around data access, permissions and vendor dependence, especially when tools connect to customer records, email, financial systems or internal documents. Owners weighing AI purely as a labor-saving tool may miss those implementation costs.
Separate research reported by Stacker and Thimble has pointed to rising AI-driven attack methods, cyber insurance gaps and the disproportionate effect of cyber incidents on small firms. Those figures should be treated as national context rather than Arizona-specific findings, and they should not be read as an extension of the Careerminds survey.
The practical point is that AI savings should be evaluated alongside security exposure. A tool that reduces administrative work may still require stronger access controls, multi-factor authentication, employee training, cyber insurance review and closer monitoring of data flows. Those issues overlap with broader AI agent security risks and with general AI cybersecurity protection guidance for small businesses.
Questions to answer before automating a role
Owners considering whether to replace or supplement a job with AI can start by separating task automation from role elimination. The Careerminds findings suggest several practical questions:
- Identify the work being automated. Is the AI tool handling a narrow task, a repeatable workflow or most of a person’s role?
- Assess what the employee knows. Consider customer relationships, informal process knowledge and company-specific judgment that may not transfer to software.
- Estimate the human effect. If a role is reduced, consider whether the worker can be retrained, reassigned or supported through a transition.
- Review customer-facing consequences. A tool may complete routine work quickly but perform poorly when a customer has an unusual request or expects a familiar person.
- Check data access. Determine which systems the AI tool can view, change or export before connecting it to business records.
- Price the full implementation. Include subscriptions, training, security controls, insurance requirements and review time, not just the software fee.
- Keep a human review path. Decisions affecting workers, customers, money or compliance should not become final without a responsible human decision-maker.
Survey and source limits
Careerminds has a commercial interest in workforce transition and employee development, so its framing of AI adoption as a workforce-management issue should be read with that context in mind. The Arizona-specific estimate is also drawn from a national survey without a disclosed state-level sample size or margin of error.
The cybersecurity and insurance figures referenced in the article come from separate sources and have not been independently verified against Arizona-specific data. They are useful for identifying related risk areas, but they do not measure the security exposure of the Careerminds respondents.
The evidence supports a measured conclusion: many Arizona small-business owners see meaningful AI potential inside existing jobs, but loyalty, institutional knowledge, customer relationships and cyber risk make replacement decisions more complicated than a simple payroll-versus-software comparison.
