The AI has moved from the order point to the schedule
This month's announcements are not about taking orders. They are about forecasting labour — the largest controllable cost in most franchised units, and the one a franchisee is supposed to control.
For two years the visible face of AI in franchising was the drive-thru speaker. It made for good video and easy headlines, and it was, in franchise-economics terms, a sideshow. What arrived this month is not. Across restaurants, hotels and multi-brand service groups, the systems being announced do the same thing: they forecast demand and tell an operator how many people to put on. That is a different order of consequence, because labour is the largest cost line most franchisees are told they control.
This is reference information, not legal, tax or investment advice. Figures below are published by third parties and compiled as-is, and the vendor claims are identified as such. Check the sources before acting on any of them.
Three announcements, one function
Franchising.com's 20 September survey of operational AI reports Bojangles using AI for sales forecasting and employee scheduling alongside its drive-thru assistant, and Head to Toe Brands — a 200-unit group spanning Frenchies Modern Nail Care, Bishops Cuts/Color, The Lash Lounge and Delta Crown Extensions — running a single daily dashboard that aggregates sales and performance across the portfolio (source: Franchising.com, 2026).
On 2 September, Hotel Dive reported Aimbridge Hospitality's launch of LIFT, a proprietary tool that consolidates labour data from several systems and uses AI-assisted forecasting to connect forecasting, scheduling and execution on one screen. Aimbridge's president for select service described labour as one of the hardest areas to get right but the most important in determining a hotel's profitability; the company called pilot results positive and published no figures (source: Hotel Dive, 2026).
And Nation's Restaurant News's 18 September preview of FS/TEC 2026 — more than 600 operators and 650 vendors at the Gaylord Texan in Grapevine, Texas — lists agentic-AI sessions from Oracle and Momos and a session from Harri titled around AI scheduling that learns (source: Nation's Restaurant News, 2026).
Why this is the consequential layer
A voice agent that mis-hears an order costs a franchisee a remake and a bad review. A forecasting system that staffs a Saturday wrongly costs a shift's payroll or a shift's sales, every week, compounding. The first is an incident. The second is the operating model.
The timing makes the point sharper than it would otherwise be. Meritage Hospitality Group, operator of 314 Wendy's restaurants — about 5% of that chain's US system — filed for Chapter 11 on 18 September, citing beef inflation, discounting and marketing misses, with franchise store-level margins at a 30-year low and store-level EBITDA down 48% (source: Restaurant Dive, 2026). Into margins like those, a scheduling tool is not a productivity toy. It is a lever on survival, pulled by whoever holds it.
A franchise agreement says the franchisee controls labour. A forecasting system the franchisee did not choose, cannot audit and cannot overrule quietly moves that control somewhere else.
The question nobody is asking out loud
Every one of these systems produces two outputs at once. It tells an operator what to do, and it tells whoever else has access how the operator performed against what it said. That second output is new. Field consultants have always second-guessed a manager's rota; they have never had a timestamped record of the recommendation and the deviation.
None of this is sinister, and most of it will be useful. But it changes what a compliance conversation can be about, and franchise agreements drafted before these tools existed rarely say anything about it.
What a franchisee should establish first
—Is the forecast advisory or expected? Ask for that in writing, because "a tool to help you" and "the standard you will be measured against" are different obligations.
—Who sees my deviations from the recommendation, and for how long is that record kept?
—What data does the system take from my unit, and may I get it back in a usable form if I leave the system or sell?
—If the forecast is wrong in my market and I staff to it, whose cost is that?
—Was it piloted in franchised units, or only company-owned ones? Bojangles' own deployment is 280 company stores out of 452 — the franchised experience is the smaller half of that sample.
On the vendor numbers
The money is arriving ahead of the evidence. Delightree announced $25 million in August to build what it calls an agentic operating system for franchise and multi-unit brands, across more than 6,000 locations, claiming businesses on the platform open new locations 25% faster and frontline teams reclaim up to 30% of their time (source: PR Newswire, 2026). Those are company figures, unattributed to any study or method, and we could not trace them to one. They may well be true. They are not yet evidence, and a franchisee should not pay for them as though they were.
The useful posture is neither refusal nor enthusiasm. It is insisting that a system which now sets the shift be held to the same standard as any other mandated cost: piloted where you operate, measured in numbers you can see, and reversible if it is wrong.