Allison Carpenter
Product Marketing Manager, Engage
What 30-plus restaurant leaders told us about automation, loyalty and guest data in Austin
Ask a room of restaurant marketers whether software should automatically ban suspicious loyalty
accounts, and you’ll get a fast, near-unanimous no. Ask whether it should automatically send those
same accounts a warning, and you’ll get an equally fast yes.
One attendee explained the difference better than we could have: “Anything that can be undone can
be automated.” Let software make the calls you can take back, and keep a person on the ones you can’t. We’ve been quoting it ever since.
Most of the AI conversation in restaurant tech, ours included, centers on what the models can do. The restaurant leaders we spoke to cared more about what happens when a model gets something
wrong. Their question is the better place to start, and it came up in both rooms.
Why we bring these groups together
PAR’s Voice of the Customer program gives restaurant brands a direct role in shaping what we build. Throughout the year, customers challenge early ideas, test products through betas and pilots, and work with our teams in product advisory groups, including the group that helped shape our newly launched Guest360.
We hold two in-person events each year, with virtual sessions in between. This year, 164 restaurant leaders from 98 brands have participated.
The Executive Forum brings senior leaders together to discuss the decisions facing their businesses and where restaurant technology needs to go. The User Group brings marketers, operators and technology teams into the details: how a workflow actually runs, where a guest gets stuck, what would make a product useful on a busy day.
We need both perspectives to make good product decisions. Just as importantly, we want customers to have space to learn from one another, swap ideas and talk through the opportunities and challenges they share.
At this year’s VoC Fest, more than 30 customers joined our product teams for two days. Executives debated strategy and placed poker-chip bets on where restaurant technology will be in 2030. The following day, the teams using that technology worked through early products and marked up draft concepts.
Their feedback revealed some clear expectations for AI, along with a few places where the industry may be making things more complicated than they need to be.
Automate the reversible
Operators showed us how messy “suspicious” really is. A bakery regular can check in three times a day and be completely legitimate. One brand described a flagged account that turned out to be one of its own trainers. The operator apologized personally and loaded the account with points. Several brands said much of their fraud comes from employees, which no guest-facing rule will catch.
So the line they drew was practical. Warnings, verification prompts and earning caps can fire automatically. Bans, which hit a real guest and are hard to walk back, stay with a person.
How much fraud matters depends on who you ask. Several executives said fraud isn’t a top-two priority. One put it as stopping one bad actor while turning away a hundred new guests. A day later, operators who track fraud by hand told us they’d use a better tool tomorrow. The executives are protecting growth. The operators are the ones cleaning up after it.
That gap is why we talk to both. The leaders who buy restaurant technology and the teams who use it every day see different problems, and a product built for only one of them will let the other down. Our Voice of the Customer program is set up to hear from each: executives at the Executive Forum and our Customer Advisory Board, and the marketers, operators and tech teams who live in the platform at the User Group.
Show the math, or it doesn't count
Brands want to show franchisees what an offer actually changed. Redemption counts don’t establish whether someone visited more often, spent more or would have purchased anyway.
That’s why customers kept coming back to A/B testing and actionable results. One leader’s response to a demo was blunt: “If it isn’t actionable, I don’t care.” Another’s requirement for AI was equally direct: “No hallucinations.”
The implication goes beyond reporting. If a brand can’t explain why a system recommended an offer or how it measured success, that brand will struggle to defend the decision to its franchisees.
Every automated recommendation creates an obligation to show the evidence behind it. The more responsibility AI takes on, the more consequential that evidence becomes.
Restaurants aren't Netflix
Personalization got a dose of realism. As one executive put it, you can’t build an experience around one guest’s hatred of mustard. Most restaurant menus don’t have enough variety for true one-to-one personalization.
What works is one-to-few: a handful of meaningful segments, clearly labeled, with a light personal touch on top. The best evidence came from asking guests directly. One brand found that preferences guests stated in a survey, like how much heat they want, converted at six times the rate of preferences guessed from order history.
Most brands are sitting on that opportunity. Guests will tell you what they like if you give them a reason to.
The friction nobody budgeted for
“Things we didn’t think were friction three years ago are friction now.” That line came from an executive whose drive-thru gives loyalty sign-up about 75 seconds at the speaker box.
Across the room, brands want the phone number captured first at the register, not last. They want one-tap sign-up the moment an order is confirmed, and apps that put ordering front and center, with points one tap away. (“You can’t eat points” got nods around the table.) One brand moved its web loyalty mix from 11% to 41% just by redesigning sign-up.
The best insights come from each other
Some of the best moments didn’t involve us. A brand that had tested loyalty tiers walked a table through its results while another brand, still deciding, asked the questions it needed answered. Another shared that making a sign-up reward available immediately cut the time to a second visit from about 30 days to about 10.
“Validation that we’re on the right path, facing the same issues,” one attendee wrote afterward. Another called it the best vendor event they attend and asked us never to turn it into a big conference.
That’s feedback we take seriously. The format needs to leave enough room for disagreement, follow-up questions and conversations that weren’t on the agenda.
After the encore - What we take back to the roadmap
Brands that signed up for early access are now working directly with our product managers, and the ideas raised in Austin are being scoped with the teams who own them.
We could have written an automation rule ourselves. The one we’ll actually use came from a customer. That’s what we mean when we say PAR is built with restaurant brands, not just for them.
Interested in participating?
Thank you to our VoC Fest ‘26 sponsors
