The disembodied voice at the Wendy’s drive-thru sounded to me like a new employee.
But when I told “him” I didn’t intend to use the fast-food company’s mobile app, a live female human voice took over and took my order. This was my first introduction to an AI-powered dynamic text-to-speech.
Welcome to the new ordering experience for consumers, where disembodied voices are greeting them via menu booths and over smartphones. It’s a fledgling market lorded over by SoundHound AI Inc., a publicly traded company backed by NVIDIA Corp. and partnering with Perplexity that some analysts breathlessly predict could top a $1 trillion valuation by 2050.
The staggering expectations have led many to anoint SoundHound as AI’s next “it” company.
“In the food space, staffing has been an issue. Restaurants don’t want their employees distracted by taking orders but delivering a greater guest experience,” Mike Zagorsek, chief operating officer at SoundHound, said in an interview. “AI offers upsell to customers for dessert, for example, and that AI ticket size tends to be higher than those taken by humans,” he added.
“We were AI before it was super cool,” Zagorsek boasted.
Today, SoundHound’s voice AI technology, which incorporates a large language model and ChatGPT, is available in more than 10,000 food locations in the U.S. for Chipotle, White Castle, Panda Express, Jersey Mike’s, Casey’s Pizza and others.
To get there, SoundHound has wheeled-and-dealed with a number of partnerships the past few months.
Last month, SoundHound said it bought key assets from Allset, an online-ordering platform that connects large and small restaurant chains with customers at Subway, Buffalo Wild Wings and IHOP Restaurants.
In May, it teamed with Perplexity, a developer of LLMs, to help its Chat AI bot to better answer requests across smartphones, cars and Internet of Things devices.
In March, SoundHound partnered with NVIDIA to introduce an in-vehicle voice assistant powered by a LLM on the NVIDIA Drive platform. This will offer real-time and generative AI capabilities for seamless responses without cloud connectivity. [NVIDIA has also invested significantly in SoundHound.]
So far, the moves have added up to generally positive feedback from customers, many of whom are getting used to the idea of ordering food over their smartphone or in person from a bot, Zagorsek said.
“There are people who are change averse,” he said. “But most believe the AI system is implemented well and consistent.”
‘Most Undervalued” AI Stock
It’s too early to gauge the precise impact of AI/restaurant sales on SoundHound’s total revenue, but plenty of market watchers believe that part of the company’s business will take it to new financial heights. Motley Fool has gone so far as to call SoundHound the “most undervalued” AI stock.
SoundHound’s advocates are betting their sky-high expectations on a global speech and voice-recognition market predicted to grow at a compound annual growth rate (CAGR) of 25% from 2023 to 2030, according to Fortune Business Insights.
SoundHound, which posted revenue of $11.6 million for the first three months of the year, merely needs to keep pace with the market to generate $165 million in revenue by 2030. Should it continue to grow at a CAGR of 20% from 2030 to 2050, it could generate $2.7 billion in revenue by 2050.
“The concept of an AI assistant is super important,” LogicMonitor CEO Christina Kosmowski said in an interview. “It lessens tedious tasks for employees and it introduces consumers to a simple, easy-to-use AI experience.”
Despite the flowers heaped on SoundHound, which went public by merging with a special purpose acquisition company (SPAC) in April 2022 and whose current market valuation is $1.3 billion, the company faces a raft of bruising competition from Alphabet Inc.’s Google, Microsoft Corp., and others integrating voice-recognition tools into their ecosystems. There is a silver lining of sorts for SoundHound: Many companies are still reluctant to share their data with tech behemoths.
At the same time, there is lingering trepidation from leery consumers. McDonald’s canceled its three-year experiment with AI drive-thru tests with IBM Corp. following botched orders that included hundreds of chicken nuggets, and ice cream loaded with bacon.
Still, optimism remains downright hyperbolic around SoundHound, based on its backing from NVIDIA, slew of partnerships and escalating number of fast-food joints using its technology.
Varun Grover, global lead, AI, for Veritas Technologies, a multi-cloud data resilience company, won’t go as far as to predict a $1 trillion valuation for SoundHound by 2050, but he does foresee a promising trajectory. He said the company’s expertise in multimodal AI — integrating voice, audio and contextual understanding — “sets them apart.”
“Success will depend on sustained innovation and expansion across industries,” Grover said. “While it’s a bold prediction, SoundHound’s technology and market penetration make it a company worth watching closely.”
