• AI visibility

Guide · AI search

When a new business asks AI how to take payments, how does a payment company get chosen?

By being the provider the answer names, with fees and features stated correctly, at the moment the business opens its doors. Small merchants choose a payment company mostly on cost and ease, often at founding, and tend to stay. An AI answer at that moment can decide years of card volume.

By Underneath Team Guide Updated

The short version

  1. The founding moment is the market: in a 2026 Sonary survey (opens in a new tab) of 3,603 small businesses shopping for payment processing, 44.1% were just starting their business.
  2. Cost decides the rest: among businesses with an existing setup, 35.2% named high fees as their biggest problem, and Javelin (opens in a new tab) found that lowering total cost remained the top reason for choosing a primary card processor.
  3. Volume compounds: Square processed $72.8 billion in the second quarter of 2026, up 13%, according to SiliconANGLE (opens in a new tab), and businesses on Stripe generated $1.9 trillion in 2025, up 34%.
  4. Fees are easy to misquote: in our pricing study, 61.9% of 840 software plan prices quoted by four AI assistants were fully faithful to the official page.
  5. Owners already use the tools: the U.S. Chamber of Commerce (opens in a new tab) reports that 58% of small businesses say they use generative AI.

Who chooses a payment company, and what is one merchant worth?

Usually the owner, often in the first weeks of the business, and the merchant is worth years of card volume.

A bakery, a mobile dog groomer, a yoga studio, a two-person online shop: the buyer for an all-in-one payment company is rarely a procurement team. It is the owner, deciding quickly, often alone, and often before the first sale. Sonary’s survey found that 44.1% of payment shoppers were just starting out, the single most common answer. Sonary is a comparison site that earns commissions from featured providers, so treat its survey as industry research with a commercial interest, not an independent census.

What that merchant is worth comes from processing volume, not a contract. Payment companies earn a share of every card payment, plus software, hardware and add-ons such as invoicing or loans. The public fee schedules show the scale of the take:

Small percentages on a growing business add up over years, and the choice tends to stick. Javelin reports that small businesses “widely express satisfaction” with their primary card provider, and that merchants selling both in stores and online were less likely to switch than a year earlier. Our inference: the cheapest merchant to win is the one choosing for the first time, and the hardest to win is one already settled with a rival.

The leaders show what winning early looks like. Stripe says businesses on its platform generated $1.9 trillion in total volume in 2025, that it serves more than 5 million businesses directly or through platforms, and that 25% of all Delaware corporations are now created with Stripe Atlas, its incorporation service. In Block’s second quarter, Square’s volume growth in the US accelerated to 10% while international volume grew 28%.

Where does AI already sit in a small merchant’s decision?

In the research before sign-up, where owners already use AI for daily work, and increasingly in how they sell.

There is no public survey we could find on how many owners ask an AI assistant to pick a payment provider. What is documented is that the tools are part of daily work. The U.S. Chamber of Commerce’s 2025 report found that 58% of small businesses say they use generative AI, and it describes AI use as more than double the 2023 level.

AI also touches payments from the selling side. In September 2025, OpenAI (opens in a new tab) launched Instant Checkout in ChatGPT, “powered by the Agentic Commerce Protocol, built with Stripe,” and said more than 700 million people used ChatGPT each week. Merchants paid a small fee on completed purchases. The feature’s status has since changed: Reuters reported (opens in a new tab) that OpenAI ended Instant Checkout in March 2026 to focus on product discovery and merchants’ own checkouts, FashionUnited (opens in a new tab) described it as scaled back, and OpenAI’s help page (opens in a new tab) still says ChatGPT may show an Instant Checkout option “for some eligible products and merchants.”

The detail matters less than the direction. Stripe’s 2025 letter (opens in a new tab) describes an Agentic Commerce Suite for selling across several AI interfaces and work with Microsoft on Copilot. A reasonable expectation is that owners will add a new question to their checklist: “can I sell through AI assistants with this provider?” Sellers weighing a marketplace instead of their own store face a related choice, covered in how marketplaces win sellers through AI.

What do small business owners ask AI about payments?

How to start, what it costs, what fits their trade, and whether a provider can be trusted. We wrote the sample prompts below to show how merchants tend to phrase these questions; they are not recorded queries.

StageIllustrative prompt
Getting started“How do I take card payments at a farmers market stall?”
Cost“Cheapest way to accept payments if I sell about $4,000 a month”
Comparison“Square vs Stripe vs PayPal for a mobile dog groomer”
Fit“Best payment app for a hair salon with bookings and tips”
Integration“Which payment provider works with QuickBooks and my booking tool?”
Trust“Why do payment apps freeze accounts, and which one is safest?”
Selling channels“Which payment provider lets me sell through ChatGPT?”

These questions mirror what owners say they need. In Sonary’s data, established businesses named high fees (35.2%) more than slow settlement (12.7%) and outdated equipment (8.1%) combined. Javelin also tracks which payment types small businesses accept, and found in-store buy now, pay later acceptance fell from 46% to 32% of respondents in a year. Owners ask about the methods their customers use, so an answer about “which provider accepts Apple Pay and pay-later” is a selection question too.

How does an AI answer turn into processed volume?

Through self-serve sign-up: the answer names a provider, the owner signs up, and card volume follows for years.

Most small merchants never speak to a salesperson. The path, as we infer it from how these products are sold:

  1. Question. “How should a new food truck take cards?”
  2. Answer. The assistant names two or three providers and summarizes fees and hardware.
  3. Check. The owner opens a pricing page, perhaps a review site.
  4. Sign-up. Self-serve onboarding, often within minutes.
  5. Volume. Every card payment for as long as the business stays.
  6. Attach. Point-of-sale software, invoicing, payroll or financing.

Two facts make step 2 unusually valuable. The merchant often has no incumbent to compare against, and Javelin’s data suggests that once satisfied, small businesses rarely move. A provider missing from the answer at the founding moment may not get a second look until something goes wrong.

Our guide to linking AI answers to pipeline and revenue covers how to measure that path; for payment companies the useful unit is new accounts and their first-year volume, not traffic.

What decides whether an assistant names your payment company?

Platforms document little about provider recommendations; studies point to independent coverage, consistent facts and reputation evidence.

Documented by platforms. According to Google, its AI features can handle one merchant question with “query fan-out” (opens in a new tab), running several related searches on subtopics before writing the answer. A question such as “Square vs Stripe for a salon” can therefore pull in pages about fees, salon software and reviews separately. For its shopping results, OpenAI’s help page says Instant Checkout items “are not preferred in product results.” OpenAI scaled the feature back in March 2026, according to FashionUnited, though that help page still describes it for some eligible merchants. We found no platform documentation on how assistants choose payment providers.

Observed in studies.

  • Independent coverage. Trade press and comparison sites count, not only your own pages: in our brand study, every tenfold rise in outside sites naming a brand within the cited pages came with 4.7 times the odds of a recommendation.
  • Reputation sources. An owner asking whether a provider is safe will be shown review sites: in our “is this brand legit?” study, 88.0% of answers cited a review or complaint platform, 61.7% of review-platform citations went to Trustpilot and the BBB, and 99.7% of answers raised at least one problem.
  • Price fidelity. In our pricing study, when an assistant’s price differed from the official page, 39 of 64 differing figures appeared on another page of the vendor’s own site.
  • Self-serving lists. Our study of “best of” lists found that 24.2% of 269 AI-cited numbered lists ranked their own publisher first. It is worth checking who wrote the lists cited for your own category.

Our inference for payment companies. Fee schedules with plan tiers, card-present and online rates, and fixed fees are exactly the kind of detail an answer can flatten. Old pricing pages, help articles and partner pages that state outdated rates are a likely source of wrong quotes. And the reputation questions that matter most in payments, frozen funds, account holds and reserves, are the ones review platforms carry. Sonary’s survey also advised owners to ask about “holds or rolling reserves on new accounts” before signing.

The pricing study covered software subscriptions, not payment fees, and the reputation study did not include payment companies, so these are expectations to test, not measured facts about this industry.

What does a payment company lose when AI gets its fees or fit wrong?

The founding-moment merchant, who signs up elsewhere the same day and may never revisit the choice.

We found no public data on how many merchants a payment company loses to inaccurate AI answers. The exposure follows from the facts above: owners choose on cost, many choose once, and assistants faithfully quoted 61.9% of plan prices in our software sample. An answer that quotes a provider’s highest tier as its standard rate, or omits a no-monthly-fee plan, removes it from a cost-driven shortlist without anyone noticing. Our guide on how to fix wrong brand information in AI answers explains the correction process.

What does GEO look like for a payment company?

Generative engine optimization (GEO) here means making every fact an owner checks easy to find, consistent and confirmed.

  1. One fee story everywhere. Pricing page, help center, partner listings and old announcements should state the same current rates and conditions. Retire or update stale pages.
  2. Answers by trade. Plain pages for the businesses you serve (salons, food trucks, contractors, online shops): what they need, what it costs, which hardware and which integrations.
  3. Integration facts. Which accounting, booking and ecommerce tools you work with, stated on your site and in those partners’ directories, since an owner’s question often starts from a tool they already use. Online sellers ask a parallel question about delivery, covered in how shipping platforms win small sellers.
  4. Comparison and review presence. Accurate listings on independent comparison sites and small-business publications; for how “best of” lists feed answers, see what best-of lists mean for AI recommendations.
  5. Reputation in the open. Explain holds, reserves and payout timing plainly, and answer complaints on Trustpilot and the BBB, where assistants find them.
  6. Selling-channel readiness. If you support selling through AI assistants, say exactly what, for whom and where, without overstating features that are changing.
  7. Measured visibility. Track the founding-moment questions by trade and by market; how many prompts to track explains how to size the set.

The neighboring guides cover adjacent buyers: ecommerce platforms choosing partners, and fintech software sold to banks and finance teams. Larger merchants with their own payments teams are covered in how processors reach payments teams through AI. No one can guarantee that an assistant will recommend a payment company; GEO makes the evidence it finds accurate, current and easy to verify.

What is still unknown about AI and payment-provider choice?

How often owners ask AI before signing up, and how much of a provider’s sign-ups those answers drive.

  • No direct survey. We found no public data on the share of small businesses that use an AI assistant to choose a payment provider; the Chamber figure covers AI use in general.
  • Vendor and commercial research. Sonary earns commissions and Javelin’s full report is paid; the figures here are from their public pages.
  • Studies outside payments. Our pricing, reputation and brand studies did not test payment companies specifically.
  • Changing products. Selling inside AI assistants is in flux, as the different accounts of Instant Checkout show.

Where should a payment company start?

Start by asking assistants the founding-moment questions for each trade you serve, and checking the fees they quote.

That first check shows which providers are named for a salon, a food truck or an online shop, whether your rates and plan conditions come back correctly, and which comparison, review and partner pages the answers draw on. It also shows whether account holds and payout times are described fairly.

If new merchant sign-ups and first-year processing volume are what you need to grow, talk to us about a merchant-question review. We will test how assistants answer the questions owners ask before they pick a provider, find where your fees or fit are described wrongly, and plan the content, coverage and reputation work that gives those answers better evidence. Keeping one consistent fee story and building pages for each trade you serve are part of the ongoing work described on our generative engine optimization service page.

Frequently asked questions

Do small business owners really ask ChatGPT which payment provider to use?

We found no survey measuring that directly. What is documented is broad use: 58% of small businesses told the U.S. Chamber of Commerce they use generative AI.

Why does AI sometimes quote the wrong processing fee?

Fee schedules have tiers, card-present and online rates, and fixed fees. In our software pricing study, most differing prices also appeared on another page of the vendor’s own site.

Does being in AI shopping features help a payment company win merchants?

It can become a selection factor, but the products are changing. Reuters reported that OpenAI ended Instant Checkout in March 2026 and FashionUnited described it as scaled back, while OpenAI’s help page still describes it for some eligible merchants.

Can a payment company pay to be recommended by AI assistants?

Not through organic answers. OpenAI’s help page says Instant Checkout items are not preferred in product results; recommendations draw on the evidence assistants find.

Sources

Free strategy call

Some questions are easier to answer about your own business.

Bring the one that matters most. On a free 30-minute call we’ll take a first look at it and send you a short written read afterward.

Ask about my businessHow we work