About
Built by operators, for operators.
We bring restaurant operations experience and engineering expertise together to build the tools we needed in our own stores.
Why we built it.
We run a restaurant franchise. Every store produces a POS report, a time-clock export, inventory counts, guest surveys and a set of books. They come as separate reports, and nobody puts them side by side. The accountant’s statement arrives weeks later and is built for filing taxes, not for deciding what to change. Reading all of it, for every store, every day, is a full-time job that nobody in a restaurant company has.
So we built the thing that reads it, and our own stores were the test. It got things wrong along the way. For a while it showed each period’s numbers under the wrong period’s label. Getting burned by our own tool is why it tells you when a store’s data has not arrived instead of showing a zero.
It still runs on our stores every day. If it is wrong, we are the first to pay for it.
The OpsIntel team
From our own stores
Four things we found in our own numbers.
None of this is sample data. Each one came out of our own locations while we were building OpsIntel. We have left out the names and the dollar amounts.
Our highest-labor store did not have a labor problem.
Over two periods, the books showed one location with the highest labor of the stores we compared. The crew was not the reason. A supervisor who covers several stores had their whole salary booked to that one store. Take it out and that store’s labor was ordinary.
One loss line was two others counted again.
Refunds, cancelled orders and voided orders each had their own line, and our loss total added them together. The refund line turned out to be the cancelled and voided orders again, with sales tax on top. Over the three periods we checked, our loss trend was showing nearly double the real figure.
Days with more labor hours were late just as often.
We suspected late deliveries meant thin staffing. So we compared days at the same store, on the same weekday, with similar order counts, over a year. Days averaging about four more labor hours sent deliveries out late just as often. In that comparison, a day’s total hours did not explain it.
About three in ten transfers had only one entry.
When a store lends another a case of cheese, both are supposed to record it. Over thirteen weeks, about three in ten of our transfers had an entry from only one of the two stores. Some may have been entered late or under a different item. Until both sides are entered, the two stores’ records do not match.
How we work.
We do the analysis. You run the business.
Nobody here will tell you AI runs your restaurant. Our system does the reading, the checking and the alerting, and you make the decisions.
A wrong number is worse than no number.
You make payroll and ordering decisions off these figures.
The call is yours.
Our system detects the pattern and sends you the alert: the store, the shift and the numbers behind it. What to do about it belongs to the person who knows the team.
Talk to the people who built it.
Questions about your systems, your data or whether this fits your group.