In 2019, long before the generative AI boom, McDonald’s partnered with IBM to develop an AI-based self-service system. The goal was to streamline the ordering process and reduce employee workload.
However, the project has faced many obstacles, including errors in processing orders and negative feedback from customers. According to the Associated Press, the test was unsuccessful, and McDonald’s canceled the project.
Why it matters. McDonald’s is a giant $180 billion company at the forefront of automating processes and using technology in its restaurants—far ahead of its competition.
The fact that McDonald’s is canceling this project during the investment boom and expectations for artificial intelligence is a canary in the coal mine about its limitations, no matter how many resources the entity implementing it has.
Behind the scenes. The AI system struggled with specific accents, was confused by background noise, and didn’t handle some complex orders well. As a result, it produced inaccurate and incomplete orders that frustrated customers and employees.
McDonald’s bought Apprente, a voice recognition startup, to improve its digital ordering capabilities. But it wasn’t the only AI startup the food giant bought that year. And despite the promising potential, reality proved stubbornly difficult.
Context. McDonald’s isn’t alone in exploring the possibilities of AI and automation in the restaurant industry. Wendy’s and Chipotle have also experimented with similar technologies, garnering mixed results.
The food service industry sees AI as a way to address labor shortages and improve service efficiency. However, the experience of McDonald’s reveals the difficulties and a lesson: We can't rely on AI to order food yet.
In perspective. AI has the power to revolutionize the industry, but current technology isn’t infallible and needs further refinement.
McDonald’s continues to invest in digital initiatives, such as mobile ordering and delivery, which have seen more immediate success and customer adoption.
What’s next? This decision reflects the importance of controlled testing before rolling out a new technology. On the other hand, future iterations are likely to overcome current limitations.
AI may not conquer fast food ordering in 2024, but it will certainly do so later on.
Image | Visual Karsa (Unsplash)
Related | Generative AI Creates a Gap Between Local and Cloud-Based Models. There’s Room for Both
View 0 comments