Companies nowadays are leaping at any and all opportunities to incorporate AI technology into their repertoire, which is why it’s important to look at some of the older hands on deck. McDonald’s – always on the lookout for making things faster and easier – partnered with IBM to incorporate AI technology into its operations as far back as 2021. The results were not positive, however, and McDonald’s canned the program this year. July 26th will be the final deadline for all the AI-powered drive-thru installments in the U.S. The franchise had tested an AI-based ordering system in over 100 locations across the country for close to three years before finally deciding to abandon the project.
McDonald’s AI testing program had essentially been an attempt to replace the traditional drive-thru with an AI-based ‘robot’ of sorts. Since the day it was first launched, the machine has been the source of much discussion on social media. In addition to frequent software glitches, the AI ordering system struggled to capture the customers’ spoken orders at an acceptable level of accuracy. McDonald’s expects an accuracy rate of 95% in its orders and the AI-based system struggled to meet 80% accuracy. The franchise likely stuck out with the program expecting it to improve. A viable AI-based ordering system would have helped the fast food company cut down on a significant portion of its labour costs.
But that’s not solely why McDonald’s would have stuck it out with AI. Integrating AI into voice-activated ordering systems currently depends on AI technology advancing far enough to recognise between different languages, accents, dialects, tones, the difference between meanings and so on. Needless to say, AI as we know it is nowhere at this level. The customer frustration the franchise incurred as a result of the program aside, this testing phase would have yielded a host of data that would prove to be invaluable as AI progresses further.
If the statements McDonald’s has made to the media are anything to go by, the franchise intends to continue to “evaluate long-term, scalable solutions” for a voice-activated ordering system. As it stands, there is quite a lot AI can bring into the fast food game that doesn’t involve differentiating between voices.
For one, the data the system has gathered is a valuable insight into how customer preferences for McDonald’s offerings have changed in recent years, particularly after the pandemic. AI can help incorporate this data through a customer preference predictive system that customises the fast food experience for people. This in turn can help the franchise with inventory management: the food and beverage industry as a whole is ruled by expiration and refrigeration limits. Pinpointing the inventory needs of each McDonald’s location is as near as technology can get to helping by way of slashing the food waste costs that every restaurant has to account for. Altogether, these two tools can help McDonald’s optimise their menu according to its locale. AI technology can potentially automate a continuous analysis of sales data and changing consumer trends to predict what menu items will perform well, even according to the season.
Of course, there are several potent issues behind the incorporation of AI technology into the food and beverage industry, least of all a fast food company. Chief among these complaints is the issue of the labour requirement that the technology will help ‘reduce’. These labour costs represent living and breathing humans that depend upon a company with a long history that has so far promised job security in a changing industry. Companies can promise that AI technology will only augment and not eliminate human contribution all they want, but keeping to such a promise will fly directly in the face of their promise to the company’s shareholders.
There will of course be other concerns as well. Much has been made of the role fast food companies play in affecting people’s nutrition, as well as the global noncommunicable disease (NCD) burden. Automating the decision of menu items to provide only the food items with the highest sales has the potential to drive communities into nutritional deficiencies, especially given how often people substitute cheap fast food options for their more expensive, nutritionally dense counterparts. Being flexible enough to introduce menu options that cater to certain palettes also helps McDonald’s locations stay relevant across a wide geographical spread, in the U.S. and beyond. Leaving such decisions up to AI can alienate an audience that could have been retained with some human intervention. The ever-present threat of customer privacy violation has also long predated AI.
If McDonald’s first foray into the great unknown proves anything, it’s that any step forward in uniting humanity with an artificial companion will require rigorous and thorough testing to ensure that it does not miss its mark.
(Theruni Liyanage)