AI, a Costly Ally for SMEs: Balancing Savings and Unexpected Expenses

⚡

Key Takeaways

1Many small businesses are adopting AI to reduce costs but are encountering unexpected expenses.

2Sparkles Homes used AI to overcome customs tariffs but has accrued unforeseen fees.

3AI spending per employee in small businesses is expected to nearly double by 2026.

💡Why it matters — AI promises savings, but SMEs need to manage hidden costs to avoid negative financial impacts.

The Rise of AI in Small Businesses

Artificial Intelligence (AI) is experiencing increasing adoption among small businesses, primarily due to its promises of cost savings and improved efficiency. However, this adoption is not without unexpected financial consequences. For instance, one company saw its bill rise significantly after using AI to circumvent the impacts of tariffs. Other businesses have integrated AI into their daily operations, which has sometimes led to unforeseen situations.

Take the example of Sparkles Homes, a manufacturer of decorative products for brands like HomeGoods. Faced with pressures from tariffs, the company found in AI a way to reduce its marketing costs and automate certain processes. This technology has proven crucial for the company's survival.

The founder and CEO, Brandon Lind, emphasizes that AI has allowed his five employees to keep their jobs. "AI is the reason our business is still standing," he asserts, highlighting the importance of this technology for its survival.

The Financial Impact of AI on SMEs

Using AI to reduce costs is a growing trend among small businesses. A survey by the US Chamber of Commerce indicates that 58% of small businesses were using AI in 2025, a notable increase from 23% in 2023. However, this adoption comes with unforeseen challenges.

For Amy Wood, founder of Flint Avenue Marketing in Texas, AI has been beneficial for managing websites and staff. She uses "Rachel," an AI assistant, for calls, and "Sonny," a sales assistant, for prospecting. However, these tools have sometimes generated embarrassing situations, such as an awkwardly written email from Sonny that elicited a mocking response from a prospect. Rachel, on the other hand, consistently comforted callers, even without apparent reason, requiring adjustments from Wood.

"Letting AI operate without supervision can lead to chaos," warns Wood, stressing the importance of monitoring these tools to avoid unnecessary disorder.

SMEs Adjusting to AI Costs

According to Ara Kharazian, chief economist at Ramp, small businesses, while fewer in number than large ones using AI, are doing so more intensively. On average, they spend $21 per employee on AI, compared to $11 for all businesses.

Companies with 0 to 49 employees spent $607 per employee in 2025, and plan to spend $1,034 in 2026, according to the Federal Reserve Bank of Atlanta. AI costs are now integrated into operating expenses, as explained by Bill Belknap of AEONRG, who includes them in his overhead.

Wood, for her part, sets aside $500 per month to anticipate a potential rise in AI costs. "We have invested in our processes, and a sudden price increase could jeopardize that investment," she explains.

Mark Gallegos, an accountant, notes that small businesses often inherit AI costs through the software they use, which incorporates new AI features. "Users don’t always realize the additional cost," he points out.

The Rapid Evolution of AI Adoption

The arrival of ChatGPT marked a turning point in AI adoption among small businesses. A report from JPMorgan Chase shows that companies established in 2023 took nearly two years to adopt AI, while those in 2025 took only six months. Generations Millennial, Z, and X are adopting AI at varying rates, but all show increasing interest.

Tammy Halevy of Reimagine Main Street emphasizes that active adopters of AI are enthusiastic and constantly seeking new applications for this technology. "They want more automation and use cases," she says.

The Challenges of Integrating AI

For small businesses, adopting AI represents a risk, especially with tight margins. Some are implementing safeguards to avoid excessive spending, such as limiting daily AI expenditures or training employees to manage tokens economically.

Lind learned the hard way by spending $1,000 to generate stock images, which led to the establishment of spending limits and usage guidelines for his employees.

Wood also faced unexpected costs, such as reloading tokens for her AI receptionist during a surge in calls. Despite this, she remains convinced of the savings compared to hiring human staff. "AI manages multiple roles at a lower cost," she concludes.