6 min read
Somewhere tonight, a hungry customer will type “find me a good taco place nearby that I can order from right now” into an AI assistant, and dinner will be ordered, paid for, and confirmed without a single website ever appearing on screen. No search results page. No scrolling through a delivery app. Just a conversation that ends in a transaction.
That moment stopped being hypothetical this month. On July 1, Square announced integrations that put its sellers directly inside ChatGPT and Claude, and the numbers explain why everyone is racing to be in the chat: ChatGPT alone now serves more than 900 million weekly users, and according to CNBC’s reporting, roughly one fifth of queries express direct commercial intent. In this piece, we will look at what Square actually launched, the quieter tools springing up around the same shift, and what a solo business should do about it in the next thirty days.
What Square Just Switched On
Square’s announcement describes a new ChatGPT app and Claude plugins “helping sellers get discovered and transact at the exact moment customers are making purchasing decisions.”
The first wave is food and beverage: US restaurants with an activated Square Online Ordering profile can now be discovered inside ChatGPT, where customers browse menus and place orders using Order by Cash App. Three details in the fine print matter more than the headline:
- Orders flow into the systems sellers already run. An order placed in a chat lands in the same Square Online Ordering setup, point of sale, and kitchen display a seller uses today. No new tablet on the counter.
- No marketplace commissions. Square says it does not charge additional commissions on orders placed through these integrations, a pointed contrast with delivery platforms whose fees keep squeezing restaurant margins.
- Sellers manage it from the Square Dashboard, while Square handles the plumbing: syncing business information, menu data, hours, and ordering details into a format AI assistants can use.
Square also confirmed it is working with Amazon to bring sellers into Alexa+, extending the same model into voice. “Find me a florist near the office that can do same-day pickup” is the kind of sentence that used to launch twenty minutes of tapping. Increasingly, it will just launch an order.
The Picks and Shovels Are Arriving Too
When a shift is real, an ecosystem of supporting tools shows up fast. Two launches from the same week tell you this one is real.
Lantern unveiled a platform whose entire purpose is helping brands measure and improve how their products appear inside AI shopping experiences. It deploys agents that predict how AI portals interpret your products, spot visibility problems, and push fixes across product pages and catalogs, with features like agent-ready scoring and AI visibility tracking.
Alli AI launched a WordPress plugin that serves AI crawlers a complete, pre-rendered version of each page while human visitors see the normal site, and deploys structured data across the site automatically. Translation: it makes your website legible to the machines doing the shopping on your customers’ behalf.
A whole discipline is forming around a simple question: when an AI assistant goes shopping, can it find you, understand you, and buy from you? Call it agentic commerce readiness. It is 2026’s version of getting your business onto Google Maps in 2010, and the businesses that did that early will remember how well being early paid.
Four Moves to Make While This Is Still Early
1. If you use Square, flip the switches now. Food and beverage sellers should confirm their Square Online Ordering profile is activated and their menu, hours, and item descriptions are current, then check the Square Dashboard for the AI discoverability settings. Eligible sellers are included without extra fees, so for once, being early costs nothing. Not on Square? Watch your own platform closely; Shopify, Wix, and the other majors are all moving the same direction.
2. If you run a WordPress site, test AI readability. A plugin like Alli AI is one route. Even without it, paste your homepage URL into ChatGPT or Claude and ask: “Based on this page, what does this business sell, what does it cost, and how would I buy?” If the answer is wrong or vague, a customer’s assistant is getting it wrong too. Fix the page until the machine reads it right.
3. If you sell products at any scale, audit your data like an AI would. Tools like Lantern target brands with bigger catalogs, but the principle scales down to a five-product Etsy shop: complete titles, honest descriptions, current prices, accurate availability. Assistants recommend what they can verify.
4. Keep your basics machine-checkable everywhere. Google Business Profile, your socials, and your site should agree on your hours, address, and offerings. AI assistants cross-reference sources, and contradictions quietly cost you recommendations.
Who Wins When the Chat Becomes the Checkout
It is tempting to read this as another platform land grab where small players get squeezed. The delivery app era taught that lesson: third parties inserted themselves between restaurants and customers, then charged commissions of 15 to 30 percent for the privilege.
So far, this wave looks structurally different, and Square’s no-commission stance is the tell. AI platforms need real-world inventory to make their assistants useful, and commerce platforms need their sellers to show up wherever customers are. Your interests and your platform’s interests are, for the moment, aligned. That alignment is exactly why acting early is attractive: the terms tend to be friendliest before a channel proves itself.
The honest concerns are worth naming:
- Relationship distance. When the assistant owns the conversation, a chat-placed order may tell you little about who ordered or why.
- Opaque rankings. Recommendation logic inside assistants will be as murky as search rankings ever were, and you cannot optimize what you cannot see.
- Channel dependence. Any feed of orders you do not control can change terms later.
The mitigation is the same as it always was: capture what you can (receipts, loyalty offers, a card in the bag that says “order direct next time”), and never let any single channel become your whole business.
But the biggest risk right now is not being exploited. It is being invisible. One fifth of 900 million weekly users asking commercial questions is a river of intent, and it is flowing whether your business is in it or not.
Your Thirty Day Readiness Plan
- This week: ask ChatGPT and Claude what they know about your business. Note every error, then fix the sources (your site, Google Business Profile, socials).
- This week: if you are a Square F&B seller, activate Online Ordering and confirm your menu data is clean and current.
- Within two weeks: if you run WordPress, evaluate an AI readability plugin or at minimum add proper structured data.
- Within thirty days: put one direct-relationship hook into every order that arrives from a third-party channel, so chat customers can become your customers.
Be There Before the Rush
Every era of small business has one channel shift that rewards the early and punishes the late: the web itself, then local search, then social. Commerce conducted through AI assistants has just moved from prediction to product, with real orders flowing to real sellers who did nothing more than keep their data clean and switch the feature on. The winners of this shift will not be the businesses with the biggest budgets. They will be the ones whose information was ready when the assistants came looking. SoloAITool tracks this shift every week so you do not have to. Here is the question to test yourself with tonight: if a customer asked an AI assistant to buy from a business like yours right now, would it find you or your competitor?
