Do you need to start hiring "AI people," or do the people you already have need a different job description? Owners ask us some version of that question almost every week, usually while staring at an open office manager or account executive role that has not changed on paper in five years. The 2026 labor market data gives a direct answer: most of the change is happening inside ordinary roles, not in brand-new AI titles.
Here is the short version. Indeed's analysis of AI job titles shows the number of U.S. job categories with AI written into the title more than tripled between 2022 and early 2026, and 63 percent of them are outside technology occupations. For most SMBs, the practical response is not a new AI hire. It is rewriting the roles you already staff, then training and securing the people in them accordingly.
Yes, and faster than most leaders realize. Indeed Hiring Lab's July 2026 analysis tracked standardized job titles where at least five postings in a quarter put AI in the title itself. In the U.S., that count dipped from 264 in 2022 to 159 in 2023, then climbed to 822 in the first quarter of 2026, or roughly one in every 12 job titles with meaningful posting volume. The same study found AI-touched titles were more common outside tech than inside it in five of the six markets it examined, with the U.S. leading at 63 percent.
The dip matters. A trend that falls, recovers, and then accelerates past its prior peak is a stronger signal than a single burst of hype. Europe showed the same shape on a smaller scale, with Germany, the UK, France, the Netherlands, and Spain all rising sharply after 2023.
The common belief is that AI's labor impact shows up as exotic new positions: prompt engineers, AI ethicists, machine learning specialists. Those roles exist, but they are not where most of the growth is. Indeed's researchers pointed out that many of the roles now carrying AI in the title are jobs that have existed for decades.
The examples are telling. U.S. postings included a "Physical Therapist (AI Documentation)," a "Real Estate Agent – AI Lead System Included," and an "AI Autonomous Truck Test Driver." In Germany, an HR manager posting asked for someone who can use AI in HR to increase efficiency. Indeed grouped the non-tech growth into three clusters: AI enablement roles such as account managers and operations managers expected to guide adoption, AI training and content roles, and AI instruction roles such as corporate trainers and coaches.
Employer survey data says the same thing from the other side. In Kyndryl's 2026 People Readiness Report, a study of 1,100 senior business and technology leaders across eight countries, 61 percent said their organizations have already redesigned roles around AI, while 24 percent are creating new roles focused on AI management. Redesign is outpacing creation by more than two to one. That is the headline for an SMB: the job you already have is the one that is changing.
A large enterprise can absorb a slow job-design update because it has recruiters, HR business partners, and deep applicant pools. A 40-person firm competing for the same bookkeeper or sales rep in the same local market does not have that cushion. If a competitor's posting already describes how AI fits into the role and yours does not, strong candidates will read the difference.
The bigger cost is not hiring. It is the people already in those seats. Kyndryl found that 52 percent of leaders say it has become harder to find employees with the right skills for their AI strategy, yet only about a third have fully implemented training to help employees work effectively with AI tools. When the role changes and nobody updates the expectations or the training, you get slower ramp times, uneven quality, and managers re-explaining work that should have been defined up front.
| Familiar role | What the rewritten version increasingly expects |
|---|---|
| Office or operations coordinator | Uses approved AI tools to draft, summarize, and organize work, and knows what data must never go into them |
| Account executive or sales rep | Uses AI research and drafting to prepare for client conversations, and verifies facts before they reach a customer |
| Customer service representative | Works alongside an AI assistant and knows when to escalate past it |
| Loan processor or claims handler | Understands what an AI tool checked, what it did not, and where human sign-off is required |
Nothing dramatic at first, which is exactly why it gets ignored. The role is posted, someone qualified on paper is hired, and the seat is filled. The cost appears later, in a new hire who can do the job as written but not as it is now performed, and in a manager who becomes the undocumented source of truth for how AI fits into the work.
There is also a risk most leaders miss. When a role quietly starts depending on an AI tool, the access that comes with it changes too. A real estate agent whose job includes an AI lead system now touches a database of prospect information. An HR coordinator using AI to screen resumes is now part of a process that several states regulate, which we covered in our breakdown of new state rules for AI in hiring. If the job description never changed, the access review, the onboarding checklist, and the offboarding steps probably did not change either.
That is how AI ends up running inside a role with nobody formally responsible for it. It is the same supervision gap we wrote about in who is supervising your AI intern, just spread across job descriptions instead of a single tool.
It is both, and the development side is where most of the value sits. Your current employees are already in roles that are shifting under them. Unless someone defines what AI fluency means for each role, some will figure it out on their own and others will not, and the difference will show up in who gets promoted and who gets the bigger accounts.
The U.S. Department of Labor offers a useful starting structure. Its AI Literacy Framework, released in February 2026, sets out five foundational content areas: understanding AI principles, exploring AI uses, directing AI effectively, evaluating AI outputs, and using AI responsibly. It also stresses embedding learning in the context of the industry and occupation. That gives you a shared vocabulary for writing role-specific expectations instead of one generic standard, and our team-level walkthrough of what the DOL framework means for your team shows how to apply it.
Managers are the hinge. They are the ones who translate a rewritten role into daily expectations, which is why the AI adoption gap is often a manager problem rather than a tools problem. If managers cannot describe what good AI-assisted work looks like in their function, the job description will not save you.
Treat it like any other operational change, such as a new system or regulation that alters how the work is done. Research on high-performing organizations backs this approach. In McKinsey's State of AI in 2026 survey, nearly three quarters of AI high performers said they fundamentally redesigned workflows because of AI, compared with about one quarter of other respondents. The steps below are the SMB version of that discipline.
If you want a clear baseline before you start, a structured AI readiness assessment shows where your roles, data, and security stand today.
Our work here sits where job design meets the IT environment. We help clients inventory which roles already depend on AI tools, sanctioned or not, and map what data each of those tools can reach. Then we align identity and access with the rewritten roles, so the permissions an employee holds match the work they actually do, and so a departing employee's AI tool access ends the day they leave.
We also help leaders set the guardrails that let a rewritten role work safely, from approved tool lists to the Copilot and agent controls we describe in our guide to securing Microsoft Copilot and AI agents. The business outcome is simple: roles that are described accurately, staffed by people trained for them, and backed by access controls that reflect reality. It is the practical side of the approach in our leadership playbook for AI adoption in SMBs.
The companies that come out ahead will not be the ones that invented an AI job title to look current. They will be the ones that rewrote the ordinary roles they already had, then trained and secured the people in them.
If your team is moving faster with AI than your guardrails are, start with structured training rather than another tool. AI University gives your people role-based AI training with security built into the material, not bolted on afterward.
If you would rather talk through your specific environment first, book a strategy call with Securafy and we will walk your current AI usage, exposure, and the fastest path to safe adoption.