Square Makes Squaremates Its Main U.S. Seller Dashboard AI

By Rebecca Stone

Square said on November 12, 2026 that Squaremates, its AI business agent, is generally available to U.S. sellers and has become the main way they use Square Dashboard after signing in.

The release also adds improved Square MCP (Model Context Protocol), the new Square CLI (Command-Line Interface), and Squaremates Apps. According to Square, the new tools for connecting outside AI assistants and building seller-specific applications carry no additional cost.

Square said Squaremates handles complex, multi-step work in its own secure workspace. It uses cloud compute to operate across more parts of a seller’s business, extending the agent beyond a text-only exchange.

The launch extends the operational-intelligence strategy Square began with Square AI in 2025. Its practical change is the addition of execution: Squaremates can prepare campaigns, alter catalog data, compare labor and sales, and assemble supplier orders while leaving final decisions with sellers.

For sellers, the stakes sit inside everyday operating workflows rather than in a general-purpose chatbot. An agent that can act on customer, pricing, inventory, and staffing information could alter what a business does next. Square’s announcement makes seller approval the central control.

What can Squaremates do?

Square said Squaremates builds on Managerbot, which entered open beta six months before the announcement. Managerbot monitored operations and proposed actions while the seller decided what should proceed. Squaremates carries that approval model into a wider set of multi-step tasks and finished outputs.

According to Square, a seller can ask the agent to:

• Review a sales period, identify lapsed customers, and prepare a re-engagement campaign or discount.
• Make bulk catalog changes to prices, descriptions, channel visibility, and images.
• Compare labor costs with sales and prepare a more cost-optimized staffing schedule for review.
• Identify low inventory through a beta restocking feature, prepare an order, and build it on a supplier’s website for the seller to complete.

Square’s workflows connect analysis to execution in several steps. The re-engagement path starts with a sales review, moves to lapsed-customer identification, and ends with a campaign or discount. The catalog path can update prices, descriptions, channel visibility, and images across a product set.

Square said Squaremates can change business data and prepare operational work for approval instead of stopping at recommendations. The distinction matters because a seller receives a proposed campaign, catalog update, schedule, or order ready for review, rather than instructions for making each change.

Square also said completed tasks can yield graphical sales breakdowns and fully designed, on-brand email campaigns, rather than only text responses. The company described those files as usable artifacts, giving a seller a concrete result to review and use.

Square plans another output: a daily briefing that would use permissioned email access to combine inbox content with information about the business. The announcement presented the briefing as a future feature rather than a generally available one.

How much work reaches execution still depends on seller approval and staged feature rollouts, according to Square. Those controls apply when an agent proposal becomes an action affecting customer communications, product listings, labor schedules, or inventory purchasing.

How do outside AI tools connect?

According to Square, the improved Square MCP lets a seller connect a Square account to a third-party AI assistant once. The seller can then ask questions and make changes through a defined set of Square actions.

Square said MCP is intended for businesses that already work with outside assistants and want those tools to interact directly with Square operations. It gives the assistant a defined set of account actions, while Squaremates remains available directly through Dashboard.

According to Square’s description, the connection is scoped: the seller links the account once, and later requests use the defined Square actions. That setup leaves Square’s action set as the boundary for later MCP requests.

Square’s new CLI serves a more technical audience. It gives sellers’ and developers’ own tools and agents command-line access to Square APIs for scripting, scheduling, and building on their accounts.

Square says Squaremates Apps and the CLI extend the same model to custom applications connected to a seller’s business. A seller can describe the desired application to Squaremates or use Square CLI with AI coding tools, then deploy and host the result on Square.

Square gave two example applications:

• Divide cash tips according to hours worked.
• Display live staffing information across locations.

Both examples turn operational data into a task-specific interface: one divides a payment, while the other tracks staffing. Square described these as suggested uses rather than the full set of applications sellers can build through the two paths.

What have sellers already built?

In Square’s announcement, Donnie McClanahan, who runs four cafes in the Knoxville, Tennessee area, used Square CLI and an AI coding agent to build a labor-board application. It combines schedules and timecards in one live view across his locations.

His app lets him see who is working at each cafe without checking every site separately, according to McClanahan’s account in Square’s announcement. The use case centers on a single labor view for an operator running four locations.

Fraser Thompson, who leads strategy and operations at Los Angeles-based BURGERS NEVER SAY DIE, used Squaremates to examine attachment rates and customer segments for the business’s house-made cheese sauce. Square said the analysis showed that longstanding regular customers particularly favored the sauce.

That finding prompted BURGERS NEVER SAY DIE to run social-media marketing and try to make the sauce more discoverable. Thompson said, “A few months later, we’d 12x’d our cheese sauce.” Square presented the result as a seller example in which customer analysis led to a marketing effort.

Together, the two examples cover different sides of Square’s agent strategy. McClanahan used an external development path to connect schedules and timecards, while Thompson used the in-dashboard agent to study product demand and customer preference.

What is not available yet?

Square does not currently make Squaremates available to Multi-Account Suite sellers. That leaves a defined gap in the launch despite general availability for U.S. sellers, and Square did not provide a date for Multi-Account Suite support in the announcement.

Square is also planning a Squaremates mobile app built on Buzz for team communication and task management. The planned app would include staff interaction with AI agents. Square is offering a waitlist for alpha access, not general availability.

Those limits make the rollout staged in practice. Squaremates is embedded in Dashboard for eligible U.S. sellers, seller-specific applications can be deployed and hosted on Square, and the mobile experience remains an alpha project.

Why does this matter to sellers?

Square now places AI in two operating layers. The first sits inside Dashboard, where Squaremates can inspect sales and customer data, modify catalog information, prepare campaigns, and create staffing proposals. The second sits inside a seller’s own tools through Squaremates Apps, MCP, or CLI.

The distinction gives sellers a choice about where to direct the work: inside Square’s own interface or inside an assistant or agent they already use. It also defines the boundary of each route. Squaremates works in its secure workspace, MCP exposes a defined set of Square actions, and CLI exposes Square APIs to command-line tools and agents.

The important control is approval rather than mere access. Square said seller approval and staged rollouts determine how much work moves from an agent proposal to an operational action. In re-engagement, that can mean reviewing a campaign or discount before outreach; in restocking, it can mean completing an order prepared on a supplier’s website.

Square’s announcement cites two seller outcomes. McClanahan built one live labor view across four cafes, while Thompson reported a 12x result for the cheese sauce after Squaremates analysis. The next test is whether Square can extend that execution model across more sellers while keeping approval boundaries clear and each workflow tied to current business data.

Squaremates’ next test is whether that execution model can scale without losing the approval boundaries Square has built around it.