Ted Taylor spent 25 years in the semiconductor industry before becoming an independent franchisee. He now operates three co-branded Schlotzsky's and Cinnabon restaurants across New Mexico, where he balances the customer experience against a large amount of behind-the-scenes work, including menu pricing, staffing, equipment maintenance and policies. Taylor uses OpenAI technology to get through much of that work faster. He relies on ChatGPT for time-consuming tasks such as research and drafting, and on ChatGPT Work to gather and analyze data across pricing, sales, forecasts, schedules and labor. According to the source, this reduces repricing time by 95% and saves four hours on payroll runs.
Repricing menus is an annual task that requires analyzing recipe costs, item popularity, local competitor pricing, portion sizes and the different economics of in-store and delivery orders. Taylor says ChatGPT Work brings these factors together and produces pricing recommendations in one to two hours instead of five days. He first uses it to set pricing targets based on local competitors, portion sizes and food quality, then supplies categorized menu items with current prices and specifies the increases he wants. With item-level sales history and supplier prices added, the tool can estimate sales impacts and optimize around cost-of-goods targets, returning a spreadsheet of recommended prices. Taylor said it made a deeper local comparison and helped him avoid increases that would have made some items uncompetitive.
Taylor also uses ChatGPT Work to manage labor costs. For analysis, he gave it read-only access to his sales and labor data. The tool compares scheduled, budgeted and actual hours against forecast and actual sales, reviews the previous 12 weeks, and flags overtime, overlong shifts and coverage gaps for managers to review. In a scheduling test, it received an employee roster with roles, training, availability and pay, along with staffing needs at different sales levels. Taylor and his manager compared its proposed schedule with the one she would have built, identified missing rules about adding staff during busy periods, and refined the instructions.
Source: OpenAI Blog · Summarized by HeadlinesBriefing