Who owns the AI layer — and who pays for it
Three announcements in three weeks put the same capability in three different pairs of hands: a vendor's, a franchisee's and a franchisor's. Ownership is what decides the economics.
Three things happened in the last three weeks of the northern summer, and the industry filed them separately. A venture round. A franchisee's software purchase. A drive-thru milestone. Read together, they are one story about a question franchising has not yet answered: when the AI operating layer sits between the brand and the customer, who owns it, and who pays for it?
Case one — the vendor owns it
On 28 August, Owner announced a $240 million Series D led by Growth Equity at Goldman Sachs Alternatives, valuing the Palo Alto company at $2.3 billion, with Meritech, Redpoint and Headline participating (source: PR Newswire, 2026). What the company sells matters more than the number: an AI-managed website, online ordering, mobile app, CRM, customer support and phone ordering for independent restaurants.
Set that list against what a franchisee typically receives for a technology fee. It is close to the same list. A capability that was, for two decades, a reason to buy a franchise is now available to an independent as a subscription, funded to the tune of a quarter of a billion dollars.
For twenty years the answer to "why buy a franchise?" included a technology stack the independent could not assemble. That sentence now needs a footnote.
Case two — the franchisee owns it
On 31 August, O&M Restaurant Group, a 23-unit Taco Bell operator in Oklahoma, announced it had selected Oscar AI to consolidate its store data into role-specific actions for executives, district leaders and store managers (source: GlobeNewswire, 2026).
The buyer is the interesting party. This is not a system-wide franchisor rollout; it is a multi-unit franchisee procuring its own intelligence layer on top of the brand's. That is a familiar pattern in other industries and a relatively new one here, and it raises questions most franchise agreements were not drafted to answer: whose data is it, may the franchisee run its own analytics on brand systems, and what happens if the franchisor later mandates a competing tool?
One caution, in keeping with the way this publication treats vendor material. Oscar AI's headline results — around $10,000 of annual profitability improvement per location, a 2% period-over-period gross profit improvement, a 26-second drive-thru reduction, 40 minutes a day saved on reports — are published as general customer figures. They are not attributed to a study, a methodology, or to results measured at O&M. Treat them as marketing until somebody sources them.
Case three — the franchisor owns it
Taco Bell now runs Omilia voice AI at more than 890 US drive-thru lanes across 38 states (source: Restaurant Dive, 7 July 2026). This is the franchisor-owned model in its purest form: chosen centrally, deployed to units, standardised across a network.
It is also the model with the most public scar tissue. Yum! slowed the programme in August 2025 after customer complaints and users deliberately breaking it — the 18,000-cups-of-water order became the industry's most-repeated anecdote. That the number is now 890 lanes and climbing is the genuinely useful fact, because it is the only one of these three cases that demonstrates recovery rather than launch.
What the three cases have in common
Each puts the same class of capability in a different place in the value chain, and each produces a different answer to the fee question:
—**Vendor-owned** — the franchisee rents capability directly. Cheap to start, but the brand's technology fee starts looking like a markup.
—**Franchisee-owned** — a large multi-unit operator builds an advantage the rest of the network does not have. Good for that operator; corrosive to system-wide consistency.
—**Franchisor-owned** — one standard, one negotiating position, one point of failure. Expensive, slow, and the only version that scales a fix as fast as it scales a mistake.
The adoption data says the choice is being made now
Franchisors are not weighing this in the abstract. The 2026 Annual Franchise Development Report puts chatbot use at 54% of franchisors in the $101,000–$250,000 investment band, with only 31% running a lead-scoring system (source: Franchising.com, 2026). A Franchise Insights survey reported by Franchise Times found nearly 90% of franchisors used or planned to use AI for recruitment content, but just 28.2% used it to contact or qualify leads in 2025 (source: franchisetimes.com, 2026).
Both datasets describe the same shape: broad adoption of the AI that produces, thin adoption of the AI that decides. Ownership of the deciding layer is still unclaimed — which is precisely why it is worth claiming deliberately rather than by default.
The question to put to your franchisor
Not "do you have an AI strategy?" — everyone will say yes. Ask instead: which layer do you intend to own, what will it cost me, and what happens to the tools I have already bought? A brand that can answer all three has made a decision. A brand that can only answer the first has not.
This is reference information, not legal, tax or investment advice. Figures above are published by third parties and compiled as-is; we have not independently verified the underlying data, and the vendor performance claims are explicitly unverified. Check the sources before acting on any of it.