
OnDeck's third-quarter Small Business Cash Flow Trend Report, published on 3 November, finds that 61% of the US small businesses it surveyed use AI, unchanged from the second quarter and up from 52% a year earlier. Among owners who report a benefit, improved marketing and fewer errors are the most common gains; increased revenue is cited by 19.1%. Separate payment data from Ramp's AI Index shows the median firm paying for AI spent $15.70 per employee in September; Ramp's lead economist discussed the slowdown in AI spending with TechCrunch on 6 November.
What the OnDeck survey found
The report from OnDeck and Ocrolus is based on 860 OnDeck customers with working capital loans, surveyed between 16 and 28 September. OnDeck says these businesses generally have fewer than 30 employees and under $10 million in revenue. It says AI use "held at 61% in Q3, consistent with Q2 and up nine points from a year ago (52%)."
The full results show where that use goes. Among owners using AI, 62.0% use it for marketing such as content, emails and copy, 47.2% to search for business solutions, 35.3% to automate tasks, 25.7% for budgeting or accounting and 14.6% for chatbots or customer service. Of those who search for information with AI tools, 82.9% use ChatGPT, 32.7% Gemini and 29.2% Claude.
Asked about measurable impact, AI users answered: 30.9% reported a significant positive impact, 32.6% moderate and 26.9% slight; 9.0% saw no measurable impact. Those reporting a benefit most often named improved marketing performance (39.3%) and fewer errors or less rework (39.1%), then employee productivity (31.8%), faster customer responses (27.6%), lower operating costs (26.8%) and operating with fewer staff or avoiding new hires (25.5%). Increased revenue was the least-cited of the named benefits, at 19.1%.
Looking ahead, 32.6% plan to increase AI investment over the next six months, 62.4% to keep it the same and 5.0% to cut it.
What Ramp’s payment data adds
Surveys record what owners say. The Ramp AI Index measures payments: it tracks corporate card, invoice and bank transfer (ACH) payments to AI providers across more than 85,000 US businesses on Ramp's platform. In September, 55.9% of those businesses paid for AI, and 50.0% of small businesses. The median firm paying for AI spent $15.70 per employee per month; the top 1% spent a median of $9,567.53.
TechCrunch, citing Ramp data, reported on 6 November that total spending on AI tokens, the units AI providers bill for, grew only 1.5% across August and September even as usage rose; it picked up again in October, rising 8.9% between 23 September and 23 October. Ramp's lead economist Ara Kharazian told TechCrunch the headwinds on spending growth remain strong. He linked the slowdown to price cuts and more careful buyers, and said that "it's not necessarily true that a doctor's office or a car dealership will need to spend that much on AI." In his October analysis he writes that, based on what he has heard from businesses, "the customer is in the favorable negotiating position" on the contracts OpenAI and Anthropic offer.
Why the two numbers differ
The surveys and the payment data do not measure the same thing. OnDeck puts AI use in accommodation and food services at 63%. Ramp puts paid AI adoption in the same sector at 33.0%. The samples differ, too: OnDeck's respondents are its own borrowers, and Ramp's are firms using its corporate card and bill pay platform.
Ramp gives one reason for the gap itself: its figures likely underestimate adoption because many businesses use free AI tools or employees' personal accounts, where no business payment appears. Evoogen's reading is that some of the gap is likely use on free or personal accounts. That is cheap, but it often means the company has no control over what data goes into the tool.
Both sources also carry limits. OnDeck states it has not verified the survey responses and that customers received an incentive to take part. Ramp covers US firms on its own platform. Neither covers the UK or EU, so treat the figures as a signal about small firms, not a census.
What a small business can take from this
A worked example, using Ramp's median and our own assumptions: an eight-person café paying for AI at the median rate of $15.70 per employee would spend about $125.60 a month. That is a hypothetical figure in US dollars, not a quote for any tool, and many teams pay per seat for only some staff. Ramp's median covers all firms on its platform that pay for AI, not cafés specifically.
The more useful lesson is where owners report results. Better marketing performance and fewer errors are the most common benefits; more revenue is the least common of the named benefits. Before adding tools, pick one job where the outcome can be measured, such as replies sent within an hour or invoices without rework, and check it after a month. If a tool shows no change there after a month, a cheaper seat will not fix that.
Falling prices also help small buyers. With token prices dropping, according to Ramp, it is reasonable to ask suppliers what you pay for usage, what happens when you exceed it, and whether a cheaper model would do the same job.
If you want AI tied to one measurable job in your business, such as answering enquiries or cutting rework, Evoogen can map the process, set it up and track the result. Talk to Evoogen about where AI would pay.
Sources
Evoogen checked these sources on 9 November 2026. Company and research claims stay attributed to their source; Evoogen analysis is labelled.
- TechCrunch: Ramp data shows AI spending stalling just as Anthropic nears its IPO (6 November 2026) — checked 2026-11-09
- Enova / OnDeck press release: New Report: 93% of Small Businesses Expect Growth as AI Delivers Measurable Results (3 November 2026) — checked 2026-11-09
- OnDeck: Small Business Cash Flow Trend Report, Q3 2026 — checked 2026-11-09
- OnDeck: Q3 2026 Small Business Cash Flow Trend Report, full results (PDF) — checked 2026-11-09
- Ramp: Ramp AI Index (data through September 2026) — checked 2026-11-09
- Ramp: Ramp AI Index, October 2026 analysis (Ara Kharazian) — checked 2026-11-09