Just as the restaurant industry has been transformed by technology, the annual Restaurant Leadership Conference has increasingly become a tech event. The gathering of industry execs covers all sorts of topics, but many of the big sponsors, vendor booths and breakout sessions are tech-focused, and artificial intelligence was, of course, one of the big topics of conversation this year.
We spent three days at RLC chatting with operators and vendors, strolling the trade show floor and listening to panel discussions. Here’s a handful of observations.
Restaurants need to fail faster
Conventional wisdom says that action requires deliberation. Before launching a new menu item or loyalty update, restaurants need to test, learn, check the numbers and test some more.
That might have to change. Several leaders at RLC suggested that restaurants need to be more nimble when making decisions, or risk being left behind in a fast-moving, high-tech world.
“I think what's most important, alongside all the things we've always focused on, is to be able to win in an environment where you wake up one day, and it's different from what you saw yesterday,” said Shannon Hennessy, CEO of The Habit Burger Grill.
In other words, restaurants need to fail fast and learn from it.
Amanda Tingstrom, senior customer success manager for loyalty provider Thanx, echoed this idea. She said Thanx’s most successful clients right now are the ones willing to make changes quickly, whether it’s trying a new offer, menu item or program structure.
This doesn’t mean operators should just throw the proverbial spaghetti at the wall. Every company should have a North Star that guides their decision-making, Hennessy said. “But what I've learned in this environment is that strategy has to be paired with speed.”
AI has a learning curve
As more restaurants adopt AI, they’re discovering that it’s not always easy to get employees accustomed to their new “co-worker.”
One example: Training staff when to step in and help a drive-thru voice bot, and when to leave it alone.
“I think that's maybe the most important question to ask,” said Brian Cassidy, owner of SDC Restaurants, a Carl’s Jr. and Dave’s Hot Chicken franchisee.
At his locations, experienced employees who are monitoring the AI via headset have a tendency to intervene a lot, he said. This prevents the AI from learning and hurts the customer experience. SDC spent some time addressing the issue, and now some employees are so hands-off with the AI that they’ll wear their headsets around their necks. (Apparently they can still hear when the bot gets stuck.)
Still, “there’s some stores that no matter what happens, they want to hear and they want to jump in,” Cassidy said.
Using AI is both an art and a science, and more work on the art part may be needed.
Voice AI is proliferating
Speaking of voice AI, the market for that technology is far from settled. More than half a dozen voice AI suppliers were in the house at RLC this year, several of which we had not heard of before.
This is generally a good thing for restaurants, as it gives them more choices and opportunities to test the still-developing technology. Tim Newton, CTO of Long John Silver’s, said his chain is currently in pilot mode with four different vendors.
AI voice adoption is creeping up. Dairy Queen recently announced plans to roll it out to more locations, and Taco Bell President Meg Farren revealed at RLC that drive-thru AI is now in about 600 units, or more than 7% of the system.
At the same time, no one supplier has emerged as a clear winner.
“Every drive-thru in America in the next five years is gonna have AI,” said Cassidy of SDC. “I don't know [which vendor] it’s going to be, but it's going to be someone.”
Who will foot the AI bill?
One reason AI has been able to spread so fast is that the actual cost of the technology is being subsidized by investors and AI companies.
At some point, those stakeholders will want to see a return on their investment, which means AI prices may have to go up.
This is actually already happening, said Amir Hudda, CEO of tech supplier Qu. His team conducted an internal study and found that the cost of certain AI actions is now 2 to 2.5 times higher than it was a year ago.
The ROI on AI is already pretty murky for restaurants. If costs go up significantly, it could become a downright liability.
Qu's strategy to address these costs with edge computing—basically a physical alternative to the cloud that sits inside the restaurant. Qu’s system is designed to allow restaurants to run some of their AI on the edge, which is more cost-effective than ping-ponging between the store and the cloud.
Of course, there are some tasks that make sense to do in the cloud, like analyzing data across multiple locations. Qu’s new Intelligent Commerce Platform is designed to automatically determine what should be handled locally and what should be sent to the cloud.
A Nvidia Jetson edge device offered by Qu. | Photo by Joe Guszkowski
Restaurants are data companies
Amid all the AI chatter, an equally important discussion was happening around the importance of data—using it, owning it and making sure it's integrated and easily accessible to AI.
A restaurant could have the best marketing in the world, but if it doesn’t have good data on who is visiting, it can be difficult to bring them back, said Mike Nettles, CTO of Zaxbys.
“If you want to grow, you have to know your customer,” he said.
The most effective way to do this, he said, is with a loyalty program (which explains why so many restaurants are adding or revamping them right now). These programs allow brands to gather information on their customers in exchange for free food or other perks. They can then start distinguishing high-value customers from those who are less frequent, which can inform all kinds of other decisions, like where to open a new restaurant.
Thinking like a data company also means choosing vendors that will cooperate with one another and allow a restaurant’s data to move freely across multiple systems.
“Whoever you end up choosing should already have a unified integration strategy and be willing to meet you where you are,” Nettles said. This is key because integrated data is table stakes for AI to be effective.
Inaccuracy is inconvenience
Customers love convenience, and we often think of that as a function of speed and accessibility, i.e., how easy it is to order from this restaurant? But order accuracy also plays a role in how convenient an experience feels.
At Marco’s Pizza, most customers eat their pizza somewhere other than in the restaurant. That means if an order is off, there’s no way to immediately correct it.
“You can’t get it wrong. We have to get it right,” said COO John Meyers. “It’s such an inconvenience for somebody to have to change that.”
Last year, Marco’s made a concerted effort to improve accuracy and was able to boost its accuracy scores by 8 points. “We’re seeing that now proliferate through the rest of the business, all the trickle-down effects,” Meyers said.