RTN's AI Share Group: What You Missed
On Tuesday, May 9, I hosted the Restaurant Technology Network’s AI Share Group event, where we brought restaurant technology stakeholders together to share real-world installations and insights, showcasing how AI is helping restaurants super-charge efficiencies, solve challenges and enhance customer journeys.
We kicked off the call by listening in on this Bloomberg interview with Presto Automation Chairman and Interim CEO Krishna Gupta. “It (drive-thru AI) actually ends up being one of the most immediately actionable applications of generative AI in the enterprise… I don’t think in three years there’s going to be a single drive-thru having a human take your orders. I think voicebots, whether it’s ours or someone else’s, will be pervasive, they will never get tired of delivering perfect service, upselling the customer and ultimately delivering a lower-cost, higher-revenue experience to customers.”
Nick Belsito, CEO, OpenCity, echoed this sentiment. “I definitely see zero drive-thrus or even telephone calls being taken from a human being… Just look at this call. I think there might be five to ten different companies doing voice or conversational commerce; it's 100%, the future of the industry.”
SoundHound is working with fully-corporate-owned brand White Castle exclusively in the drive-thru. “We're seeing phenomenal rates of completion and order accuracy there,” shared Mike Lauricella, Head of Channel Partnerships. The thing he’s most proud of? True end-to-end AI, with no humans in the loop.
Other players in the voicebot drive-thru arena on the call included Valyant AI and VOICEplug AI. Each shared insights and statistics, as well as interesting real-world applications.
"We've worked hard to deliver one of the most affordable and efficient AIs on the market, one that can take orders in under 50 seconds, achieve accuracy rates in the 90th percentile, and consistently deliver 3-4 times more upsell offers than average,” Rob Carpenter, CEO and Founder of Valyant AI shared with me after our Share Group call wrapped. The company has over 30 Hardee's and Carl's Jr. locations already under contract.
OpenCity’s Belsito said the company got 2 million views on a Carl's Jr.video the day they launched. “We're averaging 10-15 hours of labor savings, and average revenue increases in the multi-unit percentages, so significant aspects from that standpoint,” he added.
One resounding sentiment around AI is that it’s got to work right out of the gate; the suppliers on our call seemed to be taking an intentional approach to “training the AI” before it interfaces with customers.
Jay Ruparel, Co-Founder & CEO, VOICEplug AI says that every customer case is a bit different; there’s no cookie-cutter approach to deploying AI for his brand. At Olga’s Kitchen, a brand that’s been around for 50+ years, many of the brand’s senior customers order by placing a pick-up order via phone call in the car. “Our approach in general has been to start at a much higher threshold of accuracy when we go live, and we do that by using real recordings between the brand and the customers to build the AI.” He also shared that the company doesn’t use synthetic or generic data to build the AI; each AI is customized to fit the unique needs of each brand.
Sign Up Now for the Next RTN Share Group
June 20, 2023 at 11am CT. Register here
Another hot category is AI-assisted staffing, which was brought to life on the big stage at MURTEC in March, where RTN’s Start-Up Alley crowned Landed the winner of the intense competition. Landed’s Founder & CEO, Vivian Wang, joined the conversation.
“We have hundreds of thousands of messages that fly through the air over text message between a candidate and an employer every day. We also leverage ChatGPT to introduce more tonality to the AI,” she said.
Landed handles sourcing, vetting, engaging in the interview set up, and engaging with candidates after they apply. The key to Wang’s approach is speed and brand voice. “On average, it takes a general manager anywhere from like 3-4 days to get back to a candidate. We get back to candidates immediately, using conversational AI, and we do so, taking on the persona of the manager,” she shared.
Self-professed new kid on the block OhWaiter alerts staff of a customer's needs without having to flag them down, and can fulfill requests by placing orders directly into the POS.
“We're a bunch of restaurant owners who had tech backgrounds in Santa Monica and built a solution for ourselves,” commented AJ Varnet, the company’s COO. “We started out texting on-prem to ease labor constraints… Through our language model, we started increasing checks by close to 30%.” The company’s newly landed partnership with It'saCheckmate is allowing them to scale, and expand into the take-out space via text-to-order.
AI In the Kitchen
ClearCOGS was a very early adopter of ChatGPT in its automated food prep and inventory forecasting solution. “We've benefited greatly from things like ChatGPT, just being able to take all of that machine learning-based forecasting and turn it into actionable answers,” commented Matt Wampler, CEO & Co-Founder.
Shaun Shankel, CEO of Fresh Technology, shared some impressive stats his company is experiencing when it comes to smart kitchen and kitchen display systems (KDS). The brand has the advantage of tapping millions of bump data from customers, to train the AI with ticket times, station times, item modifier combinations, day parts, day of the week, on a location by location basis - all to help restaurants predict promise times with more accuracy.
“Restaurants get them (promise times) right historically about 36% of the time, plus or minus 5 minutes. Coming out of the gate, the AI is getting it right 74% of the time,” he said. The guest receives a better estimate of when the order will be ready and Fresh Technology is using the AI smooth out capacity, predicting first available pick up for delivery times, deploying it via SMS. The company is building into its KDS and sharing via open API with its partners.
Nabeel Alamgir, CEO and Co-Founder of Lunchbox, made waves when he launched an internal hackathon with his engineers and came up with an AI-generated menu photography engine. “We noticed when there are pictures on an item, it’s four times more likely to get picked versus an item without photos,” he said. He also shared his vision for iteration, whereby you can look at your ingredient mix and ask, what you should build with an ingredient mix and how you should prioritize items to build a more profitable digital storefront.
AI-Assisted Language Translation
Jordan Orlick, CEO of LetzChat, pointed out the need for real-time translation services in the restaurant industry. The company boasts patented language technology that makes it seamless for guests and employees to communicate in up to 104 different languages and over 100 dialects. Just check out this impressive example of LetzChat in action via LinkedIn post by Brian Klinger, Sr. Enterprise Solutions Account Executive at Comcast Business, a strategic investor in LetzChat.
AI Ethical Considerations
We ended the call by touching on ethical considerations associated with AI. We invited Bradley Metrock, CEO. Project Voice, and Jon Stine, Executive Director, Open Voice Network, to offer predictions and prescriptions for being good citizens of AI in our particular industry.
“When you start to get into the ethics of it all, it just makes the conversation a whole lot more difficult… This call is a fascinating one to ask, who is trustworthy in this space? And why?” commented Metrock.
“There is a human behind every one of your applications,” commented Stine. “There is a human making those decisions. Are we accountable? Have we thought through the ethical issues? Have we thought through the potential implications?”
As the call came to an end, there was a lot to ponder. And a lot to look forward to, as we plan the next AI Share Group, where there will likely be a host of new AI players and use cases to explore.
Join the conversation on LinkedIn.
Visit restauranttechnologynetwork.com for more information about membership.
WATCH the special highlight reel from the RTN AI Share Group below.