Implementing Voice AI: What Restaurants Must Know
When Yum! Brands recently announced plans to expand voice AI technology across Taco Bell drive-thru locations — aiming to implement voice AI in hundreds of stores by the end of 2024 — HT reacted with a series of questions we felt all restaurant operators should be asking about voice. It started as a conversation in my LinkedIn DMs, and soon our MURTEC Content Council was weighing in.
Here, we present highlights from this lively discussion of voice AI’s opportunities, potential downsides, and best practices.
How does voice AI integrate with POS, menu boards, and loyalty?
“It will be imperative for voice AI to integrate with the core systems used in restaurants,” notes Daniel J. Connolly, PhD, Dean of the School of Business, St. John Fisher University. “It is increasingly becoming a common way for people to interface with technology. Just look at the popularity of Siri and Alexa. Consumers and employees will want this same convenience.”
Robert Peterson, Area Vice President, New Business North America, Oracle Restaurants, says voice AI will become a standard user interface (UI). “It’s no different than the evolution from punched card to keyboard to mouse to touch screen,” he notes. “And, as a UI connecting digital transaction platforms and guest engagement platforms, the integrations will leverage existing architecture. I would recommend that restaurants adopt the Restaurant Technology Network’s data standards for new solutions coming to market.”
Will voice AI facilitate more efficient staffing?
Voice AI obviously has the potential to shift staff away from taking guest orders, freeing them to focus on other critical tasks, which can improve workflows. Peterson describes voice AI as “an employee that is always on, professional, cost effective, and only talks back when the brand tone allows! That is efficient staffing.”
Courtney Radke, VP Innovation, SageNet, says, “Automating routine tasks such as order taking, payment processing, and handling common customer questions frees up the staff to focus on more customer-experience focused tasks and value-add activities.” This suggests that voice say hi could actually enhance the human touch at the drive-thru. “I like to say that technology, when implemented correctly, should allow the employee to focus on the human interactions that keep the customers coming back,” says Radke. “It allows them to be true Brand Ambassadors. But.... the key is ‘when implemented correctly.’”
But Lee Holman, Lead Retail Analyst, IHL Group, cautions, “Voice AI still needs to be trained on retail models, and the voice style should ideally match the region.”
Will voice AI improve order accuracy and wait times compared with human staff?
Some surveys and industry reports suggest that voice may reduce errors and increase speed of service. “The jury is still out on this,” says Connolly. “When the technology does not recognize questions or customized requests, or things that go beyond what was programmed, customers can get frustrated. Actual people may be better positioned to respond to and recover from a service delivery failure.”
Radke notes, “I have had mixed results, especially in the earlier iterations of voice AI ordering. Recently, the speech recognition and natural language processing have gotten much better. Plus, a voice AI doesn't get flustered and can handle multiple orders simultaneously, even during the busiest of peak times.”
Are franchisees prepared to meet the cost of implementation?
“The acceptance by franchisees will be a tough nut to crack,” says Holman. “ Some will see deployment as an unneeded expense rather than a long-term investment.”
Peterson says, “I believe the same question was asked 10 years ago regarding online/mobile ordering. But brands must innovate or fall behind. The guest will continue to increase their expectations regarding how to decide when and where to order & enjoy.”
Radke says, “With voice AI, the buzz is in high gear and the demand signals are everywhere (labor shortages, wage increases, revenue misses, etc) so the ROI should be easier to visualize. I would just recommend that restaurants remember all the other foundational technology investments needed to support these new and growing technologies.”
What new data privacy and security concerns does voice AI raise?
“Any application of new technology should raise concerns about data privacy and security,” says Connolly. “It is always important to assess the risks and vulnerabilities and find ways to address them.”
Peterson notes, “Some states already consider voice data to be biometric data. Data retention (and deletion) will be key to avoiding concerns. Being good custodians of data is always the best path and is always the right thing to do.”
Radke says, “Voice AI systems can capture and process sensitive customer information, including preferences, payment information, etc. Ensuring this data is securely captured, transmitted, and stored is essential. What happens when the first "closed" system is breached and customer data becomes public? Will consumers pull back and lose trust in the technology? There are already regulations such as GDPR and CCPA that have protections in place for voice recordings and transcripts that must be adhered to, but as the sheer volume of information/data surges, additional protections must be developed to ensure compliance and, ultimately, that due care is taken to protect our customers.”
What other questions should restaurants ask?
Robert Dawson, Founder, Invotar, who shared his AI expertise with attendees at the 2023 Hotel Technology Forum and 2024 MURTEC, suggests several additional questions that restaurant operators should be asking before deploying voice AI: “How will the voice AI handle different languages ? Is there a fallback process if the AI isn’t getting the order right? Can I ask for a human instead? And how will the AI address non-food order responses from the guest/customer? For example, what if I ask it what the weather is going to be tomorrow?”
Dawson, an avid proponent of AI, leaves us with a sobering reminder: “McDonald’s recently tried and abandoned its AI voice ordering system when it encountered issues with accuracy and customer satisfaction.”