Back to all posts

Industry Trends

Franchise Conference Takeaways: The AI Questions That Dominated This Season

Christian Pillat · March 27, 2026 · 5 min read

Franchise conference takeaways AI produced this season come down to a shift in the questions being asked: not what the technology is, but who owns the policy, what it costs per location, what to tell franchisees already using it, and who is accountable when an answer is wrong.

Convention season is the closest thing this industry has to a snapshot. Not because of what gets said from a stage, but because a few thousand people who normally think alone spend three days comparing notes, and by the second morning you can hear which problems are common.

This year the AI conversation sounded different from last year's, and the difference was not enthusiasm. It was vocabulary.

From curiosity to procurement

A year ago the hallway question was some version of "what are you doing with it?" — asked in the tone of someone hoping to hear that nobody was doing much. This season it was replaced by five questions that all assume the decision is coming.

  • Who owns this internally? Not who is excited about it. Whose objective it is, and what happens to their other work.
  • What does it cost per location, per month, at full network scale? Asked with a specificity that was absent last year.
  • What do we tell franchisees who are already using it? Note the tense. Nobody asked whether they were.
  • Who is accountable when it gives somebody the wrong answer? The question that ends most vendor conversations early.
  • What does it replace? Asked by people whose technology line has already grown twice.

Those are not curiosity questions. They are the questions a buyer asks, and the survey data has been pointing this way for a while: on the FRANdata and IFA franchisor survey, three in four franchisors expect to increase capital spending on technology and innovation, while 28% mentioned incorporating AI and increased automation among their plans — FRANdata's own summary is worth reading for the verb. Mentioning is not deploying. But a mention that has acquired a budget owner and a cost-per-location question has become something else.

The franchise conference takeaways AI conversations kept circling

Governance was the subject with the most airtime and the least resolution, and it kept arriving in the same disguise: a question about what to permit that was really a question about what is already happening.

Three versions came up repeatedly. What may a franchisee do with our operating manual in a tool we do not administer. What do we do about the operator who has been running the brand's marketing copy through a consumer chatbot for a year. And whether adding AI terms to the agreement at renewal reads as a standard or as a restriction.

Underneath all three is the structural problem franchising has and ordinary companies do not: the people holding your intellectual property are independent businesses, so the enterprise playbook — device policy, blocked domains, a mandatory tool — does not apply.

What has changed is that the legal chill on doing anything about it has thinned. The narrow joint-employer standard codified in late February gives franchisors more room to provide and require network systems without inheriting employment liability, which is the practical import of the NLRB joint employer rule franchise networks now operate under. Several founders raised it unprompted, and always as permission rather than as protection.

The honest state of the governance question is that most brands still have no written position, and the ones that do wrote it in the last six months.

Cost control was the question with no good answer

The second theme was money, and it was the one where the industry has the least to offer.

The problem is that AI cost behaves unlike any other line in a franchise technology budget. It is consumption-based, it varies by user and by month, and it is being incurred right now on personal cards in locations that never asked anyone. So a founder trying to plan is being asked to forecast a variable cost with no baseline and no meter.

Three sub-questions came up again and again: what a per-seat price does at ten times the user count, whether the cost lands on the franchisor or inside a fee franchisees already scrutinise, and how to know what the network is spending today.

Nobody I heard had a clean answer to the third, which is the one to solve first. You cannot control a cost you cannot see, and a technology fee increase justified by AI will be examined line by line by people who are entitled to do that.

Be careful with the numbers thrown around here, too. Adoption and savings percentages circulate at these events with no published source behind them, and a claim you repeat in a board meeting becomes yours. It is worth knowing which franchise industry statistics actually exist in a citable form before you quote one.

"Is any of it real yet?"

The third question was the most cheerful, because the answer has genuinely improved.

Document question-answering has crossed from demo to deployment at a fair number of brands — not because the models got better, but because the arguments about grounding, citation and logging got settled well enough for legal to sign. Marketing content is unremarkable now. Call and meeting summarisation moved from novelty to habit faster than anything else I watched this year.

What has not arrived is anything that changes how a shift runs. The customer-facing showcases still belong to brands with an engineering function and a capital budget, and it was noticeably easier this season to find founders willing to say that out loud rather than nod along.

Set that beside the structural picture in the IFA economic outlook 2026 — most franchised units sitting with operators who run several — and the useful reading is that the buyer for network AI is increasingly a multi-unit operator with a portfolio to run, not a headquarters with a pilot to justify.

What almost nobody asked

Two absences were louder than any of the questions, and they are what I would want at the top of next year's franchise conference takeaways AI list.

Nobody asked whether their own data is good enough to point a model at. Most of what a franchisor holds is self-reported and softer than the dashboard implies, and better analysis over worse inputs produces confident nonsense at speed. That question should precede every other one on the list, and I heard it perhaps twice.

And almost nobody asked what their franchisees want from this. The governance conversation is conducted entirely in the language of risk to the brand, while the operator at 11pm has a question and a phone, and will keep solving it with whatever answers fastest.

The mood shift is real and it is a good sign. But procurement energy pointed at the wrong problem still buys you a pilot nobody opens — and this industry has a decade of evidence for that.


Take something structural into next season's hallway conversations: IFA economic outlook 2026.

Get new posts weekly

Weekly at most. Unsubscribe any time.

Back to all articles

See this working on your own content

Bring one operations document and the questions it should answer. We will show you the answers and the citations live.

Schedule Demo