It is probably safe to say that neither employees nor customers enjoy the drive-through ordering experience at any fast-food restaurant. There are simply too many distracting variables, including ambient noise, passenger banter, accents, and slurring. Many customers yell at the nondescript, mechanical speaker box and typically have a very vague idea whether their order, repeated back to them in garble, is correct. Employees are most likely equally frustrated in trying to decipher what is coming through in regard to an order on their end. All of this misunderstanding slows things down at best and results in dissatisfied customers at worst. Drive-through ordering needs a better solution.
Adopting a Speech-to-Meaning Approach
A California-based startup called Apprente has developed a drive-through automation solution leveraging an AI-driven virtual assistant to replace the human order-taker. The solution addresses the greatest obstacle to using a virtual agent in this setting—ambient noise. “Most natural language processing solutions used for virtual assistants transcribe all of what is said and then analyze it,” said Itamar Arel, PhD and Chief Executive Officer (CEO) of Apprente, “our solution doesn’t do that, we use what we call speech to meaning.” In a recent article in QSR magazine, the technology was described like this:
Unlike conventional speech-to-text solutions that transcribe audio signals, Apprente’s patented sound-to-meaning technology leverages proprietary artificial intelligence mechanisms for learning and processing of speech signals to directly infer meaning from audio.
Given this approach, the quality of the audio does not impact Apprente’s ability to understand what is being communicated. Ambient noise is not an issue. Less sophisticated and less expensive microphones and speaker systems can be used.
Arel told Tractica that part of the reason its speech-to-meaning approach works is due to narrowing what the AI must understand. “The language and phrases used for restaurant ordering is a much more restricted domain than say the vast range of what Alexa or Google Assistant must comprehend,” said Arel.
Identifying the Needs of Fast-Food Restaurants
In speaking with quick service restaurant (QSR) companies about drive-through ordering, Apprente found a few nuggets of interesting information:
- QSRs spend approximately $100,000 per year per restaurant in staffing the drive-through ordering function.
- While saving costs by automating drive-through ordering is important, QSRs have found that drive-through ordering is the most significant bottleneck to their entire service process. Therefore, the key market driver for drive-through order automation is increasing top-line sales.
Arel said that Apprente is working with a QSR company currently and hopes to have its solution go live in the marketplace sometime in 2019.
Apprente is venture backed, having received $4.75 million in seed capital in January 2018.