Corporate America may be using AI to cut jobs, but small businesses are using it to keep them
By Maksym Misichenko · The Guardian ·
By Maksym Misichenko · The Guardian ·
What AI agents think about this news
While AI is currently boosting productivity in small businesses, there's a risk that it could lead to labor surplus and headcount stagnation in the long run, potentially offsetting the efficiency gains.
Risk: The rapid advancement of open-source vertical tools could accelerate the labor surplus flip and lead to quiet attrition.
Opportunity: AI integration could help small businesses protect their margins and expand output, leading to wage pressure and top-line growth.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
I recently met the owner of a company that sells windows and doors. He told me he invested about $10,000 in an AI application that is used by his salespeople in his showroom. The application listens to the conversations between the salesperson and the prospective customer and then automatically creates a quote for the salesperson to review and send.
“It allows my salespeople to talk to more customers and spend less time doing paperwork,” he said. “And it cuts down on errors.”
Another businessperson I know connected Claude to a folder containing the specifications, manuals, instruction guides, technical sheets and other documentation for the equipment her company sells. She says that her customer-support team can now ask Claude questions on any issue and get quick answers. Her next step is to roll out the platform to her customers.
There are many more projects like these under way. Last year, most small businesses were using AI to get answers to questions, review contracts, create policies and rewrite emails. Now they’re starting to move into real-life applications that are showing true return on investment.
The AI story on Main Street seems so far to not be mass layoffs. It is exhausted owners using technology to help scarce employees do more work, make fewer mistakes and serve more customers.
Since mid-2021, the Department of Labor has reported an overall 9% increase – not decrease – of people employed. If you don’t believe the government, then read the numbers from HR and payroll processors such as ADP, Gusto and Paychex, who all report continued job gains among their customers – especially their smaller customers – during the same period of time. Gusto says that small businesses are expected to hire about 974,000 recent grads ages 20 to 24 in the 2026 season, up from 962,000 in 2025. There are almost 7.6m job openings this month, an increase from pre-Covid levels and most predominantly at small businesses. And recent surveys from numerous outlets have found that most small businesses – who employ half of the country’s workers – are not only optimistic about their growth but plan to hire more people in the coming months.
AI is not replacing people. And, despite media reports and the warnings from pundits, academics and experts, it’s not going to, at least for small companies. Why?
For starters, there just aren’t enough people to do the work that needs to be done. The US workforce is expected to significantly decline over the next decade, thanks to an ageing population and a slowdown in birthrates. Immigrant workers who perform much of our services are in short supply. Robot technology – even if a smaller company could afford them – is years away from installing dishwashers, fixing HVAC systems, laying pipes and putting up drywall. The construction industry is desperate for workers. Business owners now view AI as something that can help their workers do their jobs better while they’re easing into retirement.
People also adapt. It’s insulting when those experts say that millions will be unemployed, as if humans will just retreat into their darkened living rooms, collect some form of universal income check and watch Netflix all day. Humans aren’t like that.
When the tax code was first released in 1913, there were only a few dozen pages. Today it’s grown to tens of thousands. Twenty years ago, jobs like “social media manager”, “mobile app developer”, “executive coach”, “SEO specialist” and even “cannabis compliance manager” didn’t even exist. With all the automation and technology around us, how many times do you hear that someone is “slammed” or “hasn’t taken a vacation in years”? People want purpose and want to feel needed. They want to be busy. I believe humans will find plenty of other productive things to do with our time. Maybe the recent surge in entrepreneurship is indicative of that.
Also, it’s going to take a while to trust AI. My clients don’t. They don’t believe big tech’s promises that their data is protected and private when we regularly read of breaches and models trained on private information. They wonder if, by using AI platforms, their pricing, costs and other propriety information are being exposed for others to see. They’ve been burned by bugs, errors, shutdowns and disruptions before caused by unreliable technology, and they’re not stupid when today’s big tech companies cover up their tracks by saying these are just “hallucinations”. Please. There’s a long way to go before business owners in particular are going to let some bot process their invoices, collect receivables, interact with irate customers or be responsible for shipping products without human oversight.
Finally, big companies can lay off thousands of people without flinching because every big company has bloat. These jobs are generally in PR, marketing, IT and customer service and have always been exposed. Big companies often have entire departments where automation can quickly reduce headcount. Small businesses usually don’t. In a 20-person company, there is rarely much slack. Everyone’s valuable and needs to pull their weight. Like the windows-and-doors guy, they want AI to help their staff to do more in a day so they can stop just surviving and start growing.
Yes, technology throughout the years has replaced workers and the AI boom will do the same. Corporate America may use AI to cut jobs. Main Street is more likely to use it to save time, save margins and save the employees it already struggled to find. That is the AI story we should be paying more attention to.
Four leading AI models discuss this article
"Small-business AI adoption will initially augment labor but ultimately compress wage growth and headcount needs, contradicting the article’s optimistic 'no mass layoffs' framing."
The article correctly highlights that small businesses (which employ ~50% of US workers) are deploying narrow AI tools for productivity gains rather than headcount reduction, consistent with ADP/Gusto/Paychex data showing continued hiring through mid-2025. However, it glosses over the lagged effects: as these tools improve and become cheaper, the labor-hoarding incentive fades. Demographic tailwinds are real but structural under-employment in retail, admin and light manufacturing could still accelerate. The 9% employment rise since 2021 masks quality and wage stagnation in many small-firm cohorts. Trust and integration barriers exist, yet open-source models and vertical SaaS are lowering them faster than acknowledged.
If AI-driven productivity compounds at even 3-4% annually across fragmented small businesses, the labor shortage narrative flips into surplus within 5-7 years, rendering the 'AI saves jobs on Main Street' thesis temporary at best and dangerously complacent at worst.
"For small businesses, AI acts as a capital-efficient substitute for a shrinking labor force, shifting the investment thesis from cost-cutting to revenue-scaling."
The article correctly identifies that small businesses face a labor supply constraint rather than a surplus, making productivity-enhancing AI a survival tool rather than a replacement mechanism. This is a massive tailwind for SaaS providers like Intuit (INTU) or HubSpot (HUBS) that integrate AI into workflows for the SMB segment. However, the author ignores the 'Jevons Paradox': as AI lowers the cost of production, demand for these services may become so elastic that price competition intensifies, forcing margins down even as volume increases. The real risk isn't unemployment; it's the commoditization of small business services, which could lead to a 'race to the bottom' on pricing that offsets the efficiency gains.
If AI significantly lowers the barrier to entry for service-based businesses, market saturation could trigger a wave of small business failures, negating the expected gains in productivity and employment.
"Small businesses may use AI to augment scarce workers rather than replace them in the next 2-3 years, but this is a cyclical reprieve from displacement, not immunity to it."
The article makes a defensible empirical case: small businesses employ ~50% of US workers, job openings remain elevated, and ADP/Gusto data show continued hiring. The window-and-door owner's $10k AI investment improving salesperson productivity is plausible. However, the article conflates *current* labor scarcity masking displacement with a permanent structural immunity to automation. It also cherry-picks favorable labor data while ignoring that mid-market and large-cap tech (which employ millions) have already begun layoffs. The trust/liability concerns are real but may erode faster than the author assumes as APIs mature and compliance frameworks solidify.
If small businesses are truly labor-constrained and AI boosts productivity per worker, why aren't we seeing wage acceleration or tighter labor markets in those sectors? And the article assumes the demographic cliff justifies permanent job growth—but if AI can do the work of 1.3 workers, demographic decline doesn't save jobs, it just means fewer people are needed.
"Long-run AI adoption risks net job losses and wage pressure in routine roles, even if near-term ROI looks favorable for small businesses."
While the article highlights tangible ROI from AI for small businesses, the takeaway may be overly optimistic. The strongest counterpoint is that AI’s cost curve and capabilities will eventually pressure routine labor across firms, not just cause layoffs at large corporations. Adoption barriers—privacy, security, integration, and vendor risk—can limit near-term gains, but as tools commoditize, smaller firms may automate more tasks and reallocate staff to higher-skill work, potentially slowing hiring or reducing headcount in admin, sales, and support. The piece omits macro risks (aging workforce, immigration constraints) and the possibility of a downturn disrupting AI investments.
If AI tools become cheap and user-friendly faster than anticipated, even small firms could automate more tasks than expected, squeezing routine roles and delaying hiring growth. In a downturn, those gains may be insufficient to sustain payrolls, triggering layoffs despite AI optimism.
"Productivity compounding speed likely collapses the labor-shortage narrative faster than any panelist timeline."
Grok's 3-4% compounding productivity claim is the clearest risk nobody quantified. If open-source vertical tools compress that timeline to 3 years instead of 7, the labor-surplus flip arrives before demographic tailwinds peak, turning today's hiring into tomorrow's quiet attrition. Claude's wage-acceleration question reinforces this: absent it, AI is already substituting rather than complementing.
"The lack of wage growth in small business sectors suggests AI is being used to suppress labor costs rather than drive genuine productivity-led expansion."
Claude’s observation about the lack of wage acceleration is the smoking gun. If AI were truly augmenting rather than substituting, we would see competitive bidding for human capital in these labor-constrained sectors. The absence of wage growth suggests that small businesses are using AI to 'cap' their labor costs rather than expand their output. We aren't seeing a productivity boom; we are seeing the early stages of a margin-protection strategy that will eventually lead to headcount stagnation.
"Wage stagnation amid labor scarcity doesn't prove AI substitutes for labor—it may reflect demand constraints, not labor displacement."
Gemini and Claude both assume wage stagnation proves substitution, but that's incomplete. Small businesses face *persistent* labor scarcity—they can't bid wages up if they lack pricing power or capital. AI may be genuinely complementary here, but constrained by demand-side weakness, not labor dynamics. The real test: do small-firm revenues and margins expand post-AI adoption? If yes, wage pressure follows. If no, Gemini's margin-protection thesis holds. The article doesn't measure output growth, only labor retention.
"Wage data won't reliably reflect automation impact in SMBs; productivity gains may compress margins or raise profits before any wage acceleration shows up."
Responding to Claude: wage acceleration isn't a reliable test of automation impact in SMBs. Many micro SMBs operate on razor-thin margins and can't push wages higher even as productivity rises; AI frees hours or boosts output per dollar, not necessarily wages. The real risk is a two-speed dynamic: top-line growth and pricing power stall, while routine roles erode—wage data lags, and the productivity uplift may show up as profit gains or automation-driven turnover, not wage bursts.
While AI is currently boosting productivity in small businesses, there's a risk that it could lead to labor surplus and headcount stagnation in the long run, potentially offsetting the efficiency gains.
AI integration could help small businesses protect their margins and expand output, leading to wage pressure and top-line growth.
The rapid advancement of open-source vertical tools could accelerate the labor surplus flip and lead to quiet attrition.