Square's AI may suggest staying open longer. What does the extra hour earn?

On October 23, Square said it had welcomed a team from Zenlytic, a company that built tools for asking business-data questions in plain language. Square says that team will help it build on Managerbot, with examples including sales trends, menu prices and weekend staffing.

This is a team announcement, not evidence that a new set of features reached every seller yesterday. It does not establish a new price, a general rollout date or measured time savings. The more useful guide to Square's direction is its separate explanation of how Managerbot is being designed.

That explanation describes a move away from an empty chat box. In Square's early research, sellers wanted relevant information brought to them rather than having to work out what to ask. Chat remains available for following up. Square also describes executing recommendations with seller approval. Those are the company's descriptions, not findings from a hands-on test for this article.

Square gives an opening-hours example: a seller may think an extra hour loses money, while Managerbot's comparison shows revenue outweighing the labor cost. That could reveal an opportunity the owner had missed.

But revenue above labor cost does not, by itself, tell the owner what the hour earns. Drinks use ingredients and packaging. Payments have fees. Staying open may move cleaning later or require more paid time than the advertised hour. Some costs change with the decision; others would be paid either way.

Square's own established restaurant-metrics guide makes the broader distinction. It includes ingredients alongside labor, and notes that labor involves benefits and payroll taxes as well as wages. It also names processing fees, utilities and other expenses. That guide predates this week's announcement; it is background on the business question, not evidence of a new Managerbot capability.

None of this proves Managerbot omits those costs. The public example simply doesn't settle it. Before acting, ask: is this comparing sales with wages, estimating the additional money left after the costs of opening longer, or assessing the whole business's profitability? Those answers can support different decisions.

If you have access to an assistant that can inspect your business figures, try a specific follow-up: ‘Show the comparable days behind this recommendation. Which costs of staying open longer are included? Which are missing? Keep recorded sales separate from your estimate of future demand.’ This is a suggested question, not a verified Managerbot command.

Comparable days matter. A busy event night can make an ordinary evening look more promising than it is. Square's older guide warns against judging performance by comparing unlike weekdays. Likewise, sales recorded before your current closing time cannot directly establish what customers would spend after it.

An answer that admits it lacks ingredient costs may still save work. You have a smaller question to finish, rather than a confident recommendation you must quietly reconstruct. If it can only answer the sales question, use it for that. You don't have to give up useful report help because it cannot yet decide your opening hours.

For an owner considering later opening, I'd start with a limited trial on suitable evenings, with staff agreement and paid closing work included. Compare the additional sales with the costs that actually changed. Note whether customers arrived later or simply shifted purchases they would have made earlier.

Choose the purpose before judging the result. Later hours might make sense as a deliberate service to regulars even with a modest financial return. Or the owner may want more income without adding another evening at work. A recommendation cannot choose between those goals on the owner's behalf.

I'd welcome an assistant that brings this question down to a few understandable figures. I'd also want it to leave me free to say: yes, we could earn more then, and we're still closing at the usual time.