McDonald's says it's ready to double its technology business

When McDonald's spent more than $ 300 million on dynamic startup of Dynamic Yield at the start of Big Data, this decision was a surprise. Follow-up should not. Today, the Golden Arches announce the acquisition of Appente, an AI voice system intended for fast food orders. It's a niche, but it just paid off.

The specific terms of the agreement have not yet been disclosed. But the synergies are at least immediately understandable. Apprentice-based artificial intelligence integrates within the relatively narrow boundaries of fast-food restaurants. As with Dynamic Yield's decision engine, which switches menu items based on what consumers want, anytime, anywhere, the ultimate goal of approval is to speed up the speed of each transaction. Anyone who has had to repeat his order to a screaming speaker knows that pain.

Apprent calls its technology "sound-to-speech" as opposed to "word-to-text". The difference, apart from the fact that it sounds good, is that, unlike many models of artificial speech intelligence, Apprente states that it does not transcribe what the client says, and then infer the meaning of that transcription . This goes directly from the speech signals to the result. "The company believes this solution is a better approach for use cases related to the customer experience, especially in noisy environments such as restaurants and public spaces or in cases where customers have tendency to use unfamiliar and poorly structured language, resulting in inaccurate word recognition, "Raúl Castañón-Martínez, senior analyst at 451 Research, recently wrote in a report on voice assistants. In other words, in environments such as driving service, where the stakes for fast food restaurants are incredibly high: the majority of the 68 million customers McDonald's serves every day never leave their car .

Because it works in menus of a menu containing a limited number of elements and according to a fairly predictable interaction paradigm, the system could be easier to make these interactions more natural that voice assistants more free. Window swaps in drive-in mode often involve accents, a challenge for even the most sophisticated vocal AI. Customizing orders may require search and pecking on the part of the operator, which can potentially slow the line. And Apprentice extolled its ability to incorporate order changes on the fly faster than its human counterpart, for example.

But does it really work? As with dynamic performance, the proof may be in the purchase. McDonald's has never been very happy with its acquisitions. Apparently he was quite pleased with the way the Apprentice voice assistant performed real world tests to take a picture of him. And unlike Dynamic Yield, which has remained a separate entity wholly owned by McDonald's, Learning will be entirely internal. This makes an acquisition just as defensive as the other; Burger King and Wendy's are going to have to create their own voice in conversational mode now. Not only that, but the Learning Team will become founding members of the newly formed McD Tech Labs, a Silicon Valley-based group that will work on faster or better technologies, at least, thanks to technology.

"Building our technological infrastructure and digital capabilities is fundamental to our Velocity growth plan and enables us to meet the growing expectations of our customers, while simplifying and making guest service more enjoyable for guests," said the director. McDonald's general, Steve Easterbrook, in a statement. , in reference to the strategic vision that he outlined in 2017.

On this last point, last fall, in a press release from Apprent, claimed that his AI "offered a more consistent and enjoyable customer service experience with virtual agents never sounding tired, bored, unhappy or angry. ". McDonald's declined to say whether its effectiveness would also reduce the number of human jobs.