McDonald's Gives up on AI Drive-Thru

McDonald's Gives up on AI Drive-Thru

In a significant shift in its technological strategy, McDonald’s has officially announced the termination of its AI-driven drive-thru ordering system , a project in collaboration with IBM started in 2021. McDonald’s “told franchisees that it’s winding down an AI drive-thru ordering partnership with IBM no later than July 26th, 2024,” McDonald’s stated. This decision will lead to the removal of the technology from over 100 restaurants where the system was being tested. McDonald’s, in its official statement, did not clarify the decision process or metrics used to stop the initiative. It did not even comment on the fact that it was a failure, or a partial success. But consumers tell a different story.

Unhappy Customers

The AI system’s glitches is a popular subject on social media platforms like TikTok. In one viral video, a customer’s interaction with the AI at a drive-thru went awry, with the screen erroneously tallying a total of 28 orders of Chicken McNuggets , amounting to hundreds of dollars. In another instance, a customer attempting to correct an order faced further complications as the AI added eight additional sweet teas . Many more such funny videos and comments have been accumulating on the internet for a while. Why did McDonald’s ignore these public feedback, which only represents a fraction of its unhappy customers.

Why Did They Persist so Long?

Why did McDonald’s take so long to discontinue a technology that was evidently not meeting expectations? According to a 2022 analysis, the IBM technology was underperforming . The report highlights the 80% accuracy, unacceptable by retail standards, and below McDonald’s expectation of 95%. Despite these early signs, McDonald’s continued with the tests, possibly hoping for improvements in the system’s accuracy and efficiency over time. The main reason may have been that McDonald’s faces a rapid increase in labor costs in the US, and saw AI as a potential solution to this threat to their business model.

Don’t Oversell, Don’t Overbuy

These incidents underscore the limitations of current AI technologies in managing dynamic and unpredictable real-world environments. They also highlight the potential risks of deploying such technologies prematurely in customer-facing roles. Full automation, especially in high-stakes, fast-paced environments like fast food drive-thrus, may not currently be the best strategy given the technology’s existing accuracy levels.

The allure of cutting-edge technology can be tempting for large corporations looking to innovate and streamline operations. However, tech companies and their clients must approach the integration of such technologies with caution. Overselling the capabilities of AI can lead to not only financial losses but also damage to a company’s credibility and customer trust.

Best Practice

The correct approach, as this case suggests, should involve training on the potential and limitations of the technology. This should be followed by a limited-scale proof of concept to validate the technology in real-world conditions before any full-scale implementation. Such measured steps can help in mitigating risks and enhancing the technology’s success rate in practical applications.

McDonald’s decision to halt its AI drive-thru experiment serves as a valuable lesson. It emphasizes the importance of rigorous testing and gradual implementation of AI technologies, especially in environments that are as unpredictable and varied as drive-thru services. As companies continue to explore AI’s potential, balancing innovation with prudence will be key to achieving sustainable and successful technological integration.

We are Here to Help

At System in Motion, we are committed to building long-term solutions and solid foundations for your Information System. We can help you optimize your Information System, generating value for your business. Contact us for any inquiry.