Square just made an aggressively practical bet: restaurants can now accept orders placed inside ChatGPT and Claude, without building or maintaining any AI integration. The mechanism runs through Square’s existing Square Online Ordering setup, and it’s already live for U.S. Food & Beverage sellers with an activated Square Online Ordering profile. The big business consequence is that these AI-driven transactions avoid the traditional marketplace commission fees that typically squeeze restaurants.

Here’s the money part, up front: Square is still charging its typical online ordering fees of 3.3% plus $0.30 per transaction for merchants on Square Plus, or 2.9% plus $0.30 per transaction for merchants on Square Premium. But the system does not add marketplace commission fees on top of that, which is the cost layer restaurants usually fight when visibility and fulfillment are bundled into delivery platforms. Square also pulls from the live Square catalog and dynamically maps items, pricing, complex modifiers, and stock availability, so autonomous agents are not supposed to surface out-of-stock inventory.

Under the hood, Square’s integration is designed to be mostly invisible to merchants. Sellers manage discoverability and operational details, like menus, operating hours, stock levels, and pricing, through the Square Dashboard they already use. When a customer prompts ChatGPT or Claude, the AI parses the real-time data provided by Square, lets the customer browse results, and then the customer can finalize checkout within the chat window via Order by Cash App. Depending on the AI tool configuration, the customer may also be redirected to the merchant’s standard online ordering landing page with the basket already populated with chosen items and modifiers.

For operators and enterprise teams that worry about traceability and “did the integration actually work the way we think?” Square built in a way to audit your digital footprint. Operators can use the “@” symbol to invoke the Order by Cash App plugin directly within ChatGPT, or connect via the Claude extension directory, to manually verify behavior. Once an order is placed, it routes into the seller’s existing operational flow, popping up on the Square Point of Sale (POS) and Kitchen Display System just like other direct online orders. Square also tags the order’s origin as an AI integration in backend reporting, so businesses can measure return on this new acquisition channel rather than treating it like a mystery box.

Why does this matter so much right now? Because restaurant margins are already under siege by third-party ordering and delivery economics, and the math gets brutal quickly. Square frames the typical squeeze as a blend of marketplace commissions and payment processing fees. The source cites DoorDash commission rates of 15% on its “Basic” delivery tier, rising to 25% for “Plus” and 30% for “Premier,” and a 6% marketplace fee even for pickup orders. Uber Eats is described as charging delivery marketplace fees from 20% on “Lite” up to 30% for premium placement, with pickup orders costing up to 10% if in-store pricing isn’t strictly validated. Grubhub is cited as taking between 5% and 20% of total order value depending on the marketing and delivery package.

Those marketplace commissions sit on top of payment processing fees that are “typically around 2.5% to 3.05% plus a fixed cent amount per order.” Square’s argument is simple and painfully familiar: independent restaurants might clear only 3% to 9% net profit on a good day, so giving up 25% to 30% commissions on a $40 digital order can effectively mean preparing food at a loss. Square’s pitch is that AI discovery can behave more like direct, first-party ordering: instead of surrendering a 30% cut to an aggregator, a restaurant discovered through an AI agent pays Square’s standard online transaction processing fee (around 2.9% plus 30 cents per transaction on a standard plan, with no monthly marketplace commission attached).

Square also distinguishes this from delivery-platform economics by how delivery is handled if it’s needed. Rather than charging a percentage that subsidizes a gig-worker driver network, Square uses a white-label dispatch network with a flat courier fee, often around $7 to $10 depending on distance. Restaurants can choose to absorb that flat delivery cost or pass it to the customer, which is a margin-protection lever that delivery marketplaces typically don’t offer as cleanly.

Square is positioning the integration as the first step in a larger “agentic commerce” strategy. The stakes are high, and Square provides industry data: more than 42% of consumers use AI tools to assist with shopping tasks like product discovery and comparison, and analysts project that agentic shoppers could drive nearly $385 billion in U.S. ecommerce spending by 2030. For small and mid-size businesses, the problem has been that they do not have the developer resources to build custom integrations for every chatbot, voice assistant, or new AI device platform that comes along. Square’s strategy is to act as a universal connective layer that gets sellers “where customers are going,” starting with ChatGPT and Claude.

Square is also already expanding beyond text chat. It announced it is actively working with Amazon to bring sellers into Alexa+ voice commerce experiences. On the standards and regulatory front, Square says it’s participating in major regulatory and standards groups, including the AAIF Agentic Commerce Working Group and the W3C Web Payments Working Group, to shape how AI agents and commerce platforms interact at scale. Square’s partnership with Google is also referenced as part of co-developing the Universal Commerce Protocol (UCP) spec. Translation: this is not only an app launch, it’s an attempt to influence the rules of the road so transactions can be orchestrated safely, consistently, and cheaply as AI agents proliferate.

For decision-makers, the strategic question is no longer “Will customers use AI to find restaurants?” The source points to behavior already underway. The question is whether the channel will behave like a first-party storefront or a new marketplace that extracts value through opaque, compounding fees. Square is clearly aiming for the former: AI-driven discovery and ordering that uses Square’s catalog and operational rails, charging standard online fees rather than marketplace commissions. If it works at scale, it pressures competitors to rethink commission-heavy models and forces boards to treat agentic commerce less like a novelty and more like a distribution shift with direct profit-and-loss consequences.