Disclosure: Some links in this article are affiliate links. We may earn a commission if you make a purchase, at no extra cost to you. Our recommendations are based on our own independent research and are not influenced by commissions. Read our full affiliate policy.

Forbes reported earlier this week that around 29% of companies that made AI-related layoffs in 2024-2025 have already rehired for similar roles. The headline is counterintuitive enough to be everywhere; the practical lesson is significant if you are a small business owner weighing whether to use AI to avoid a hire, replace a role, or shift headcount.

This guide walks through what the data actually suggests about where AI substitutes for headcount in 2026, where it falls short, what kinds of roles are coming back, and how a small business owner should think about the AI-versus-hiring decision without making either the layoff-and-regret mistake or the missed-opportunity one.

What the data actually shows

The Forbes piece and the underlying surveys point to several patterns:

  • Companies that cut roles based on early AI capability projections have been rehiring as the real-world limits of those capabilities became clear.
  • The categories most often cut and then rehired include customer support, content production, mid-level analyst roles, and certain operations functions.
  • Rehiring often comes with a different scope: AI handles the routine, the human handles judgment, exceptions, and customer escalations.
  • Companies that did selective augmentation (AI plus people) report better outcomes than companies that did substitution (AI instead of people).

None of this means AI is failing. It means the breakeven point (where AI fully substitutes for a role) is narrower than the early discourse suggested.

Where AI does substitute for hiring in 2026 SMBs

  • Pure data-entry and categorization tasks that previously needed a part-time hire.
  • Tier-1 support for FAQ-type questions, with a clear human escalation path.
  • First-draft content production (blog posts, product descriptions, social copy) reviewed and finalized by a human.
  • Basic scheduling and inbox triage.
  • Routine reporting and summarization.
  • Initial outbound personalization at small volume.

For these, AI plus a small amount of human oversight can postpone or eliminate a hire that might otherwise be necessary.

Where AI does not yet substitute reliably

  • Customer relationships that depend on trust, history, and judgment.
  • Complex sales with long cycles and multiple stakeholders.
  • Anything requiring physical presence (installation, repair, in-person service).
  • Quality control and editorial judgment on customer-facing output.
  • Exception handling when standard workflows do not apply.
  • Strategic decisions and prioritization within the business.
  • Brand-sensitive communications (PR, crisis response, sensitive customer issues).
  • Cross-functional coordination that requires reading context, relationships, and politics.

Where companies cut and then rehired, the work being recovered usually falls into these areas.

How a small business owner should think about the decision

1. Separate the role from the tasks

“I need a marketing person” is a role question. The right starting point is: what tasks need doing, how often, with what scope and judgment? Many roles include some tasks that AI can do well, some that need a human, and some that are mixed.

2. Map current and projected tasks against AI capability

For each task, ask: can AI do this reliably today, with what human oversight, at what cost? Some tasks are clear AI wins. Some are clear human work. The mixed ones are where most of the actual decision lives.

3. Run a 60-day trial before committing

If you are debating between hiring and AI substitution, do a 60-day trial with the AI approach first. Track quality, customer outcomes, and the actual time you spend overseeing it. Hire if the trial fails on any of those.

4. Hire for judgment, augment with AI for execution

The pattern that holds up: hire people for the parts of the role that need judgment and relationships, equip them with AI to handle the execution they would otherwise have spent time on. This often means hiring slightly more senior than you previously would have, and asking that person to operate with AI use rather than do everything themselves.

5. Avoid the “AI saves us money” framing alone

If AI substitution is genuinely cheaper for a clear-scope task, fine. If it requires extensive owner oversight to keep working, the savings often disappear. Cost is only one factor; reliability and quality matter more for customer-facing work.

Common mistakes

Cutting first, testing later

Companies that rehired after AI layoffs often discovered limitations after the cuts. Test substitution before reducing headcount.

Substituting where augmenting was better

Replacing a content manager with AI usually produces lower-quality output than keeping the content manager and giving them AI tools. Augmentation is the more common right answer for skilled roles.

Underestimating oversight cost

Every AI substitution requires some human oversight. If your only available overseer is the owner, the time cost can outweigh the savings.

Overestimating customer tolerance for AI-only interactions

Some customers are happy with AI-handled service for simple needs. Many are not, especially for higher-stakes interactions. Know your customer base before substituting heavily on customer-facing roles.

Hiring late because “AI will handle it”

The other failure mode: deferring a hire because of optimistic AI projections, then operating short-staffed and burning out the team while waiting for capability that has not arrived. If you need help now, get it.

What to do this quarter

  1. List the next 3 hires you have been considering or deferring.
  2. For each, list the tasks in scope.
  3. Mark each task as: clear AI win, clear human work, or mixed.
  4. If a role is >70% AI-doable today, run a 60-day AI-substitution trial before hiring.
  5. If a role is mostly human work with some AI-augmentable tasks, hire and equip with AI tooling.
  6. For roles already partially handled by AI, audit quality and customer outcomes quarterly.

Tools and platforms that fit the augmentation pattern

If you are equipping a hire (existing or new) with AI tooling rather than substituting outright, two existing guides on Apex Business Tech are useful starting points:

FAQ

Are SMBs making the same AI-and-rehire mistake as enterprises?

Less often, because SMBs have less headcount to cut in the first place. The more common SMB mistake is deferring necessary hires based on optimistic AI projections.

How do I know if a role is safe to substitute?

If the work is well-defined, repetitive, and has clear quality criteria, AI substitution is more feasible. If it requires judgment, relationships, or exception handling, augmentation is the better pattern.

Should I hire more senior people who can use AI well?

Generally yes. Senior people can absorb AI tools to expand their effective output, where junior people often need closer supervision to use AI well.

Will rehires command higher salaries?

In some cases, yes, particularly when the rehired roles include AI-tool oversight responsibilities. Factor this into the substitution cost analysis.

How do I handle customers who notice AI-handled interactions?

Be transparent. AI-handled tier-1 with clear human escalation is generally accepted; pretending AI is human, or failing to escalate cleanly, is what damages trust.

What about contractors and freelancers as a middle ground?

Often the right answer for variable-load work. Contractors and freelancers using AI tools can deliver a meaningful share of skilled work without a full-time hire commitment.

More from the ABT AI for Small Business Series

Bottom line

The 29% rehiring rate is a useful corrective to the early “AI replaces workers” framing. Around two-thirds of AI-related substitutions have stuck; around a third have not. The pattern of which ones reverse is clear: roles with high judgment, relationship, and exception-handling components are harder to replace than the early discourse assumed.

For a small business in 2026, the practical play is not “AI or hire” as a binary. It is task-by-task analysis, trial substitutions for clearly AI-doable scopes, augmentation for skilled roles, and willingness to hire when the work needs judgment that AI does not yet reliably bring. Owners who think in tasks rather than roles will avoid both the layoff-then-regret mistake and the missed-hire mistake, and end up with a leaner, better-equipped team than either pure substitution or pure traditional hiring would produce.