Given how important speed and accuracy are for the mountains of repetitive tasks that occur within The Golden Arches empire, McDonald's foray into AI-powered drive-throughs seemed like a recipe for success to me.
But, Ronald’s ambitious AI experiment has hit some unexpected snags and forced a retreat of their AI strategy.
By examining what's gone on here for McDonald's, I aim to pick out 3 major insights for business leaders who are navigating their own complexities of AI implementation.
The McDonald's AI experiment: A case study
McDonald's venture into AI-driven ordering began with high hopes...
The idea was simple, yet innovative: use artificial intelligence to take and process drive-through orders, potentially streamlining operations and reducing wait times.
The system was designed to understand and interpret customer orders, promising an idyllic (albeit grease-soaked) future where machines could handle the complexities of human conversation and interaction.
However, reality served up a different menu.
The AI repeatedly stumbled when challenged by the full gamut of accents and dialects that make up the demographic of McDonald's diverse customer base.
Order accuracy took a hit, and customers often found themselves repeating orders or dealing with mistakes (one customer was asked if he'd like bacon with his icecream!)... a far cry from the customer service dream that the AI experiment promised.
The backlash was pretty swift, and sharp.
Social media swarmed with customer complaints (watch here as two customers snort and giggle as the AI racks up 260 McNuggets on their order!), turning what was meant to be a tech triumph into a PR nightmare.
What's the first lesson here?
In the pursuit of innovation, McDonald's had actually compromised the thing at the heart of their business - customer satisfaction.
In response, McDonald's has decided to wind down this particular AI experiment.
(But they're not abandoning the technology altogether. Instead, they've moved on from IBM as their AI provider and pivoted to a new partnership with Google Cloud; a shift in their AI strategy rather than a full retreat.)
Start small... scale smart
The 2nd important lesson I'll point out from McDonald's AI experiment is that when it comes to AI, it's almost always wise to “start small and scale smart”.
Pilot projects allow us to test the waters without risking reputation in the eyes of an entire customer base.
I personally advise that businesses should view AI adoption as a journey of incremental improvements. Each small success builds confidence, generates insights, and paves the way for larger implementations.
"Gawd, I'm so McBored"
Prioritise internal AI applications
Another important takeaway I’d glean from the McDonald's AI story is the benefit of prioritising internal AI applications over external applications in the early days of a company's AI deployments.
While customer-facing AI might seem like the fastest route to visible results, it also carries the highest risks.
Internal applications on the other hand – AI workflows designed to augment or automate processes that assist and affect the employees and processes of a company – offer a safer testing ground.
Internal AI applications allow companies to test and benefit from the power of AI, while minimising potential negative impacts on customer experience.
For instance, AI can be used to optimise supply chain management, streamline internal communications, or enhance data analysis for decision-making.
These are much less risky first deployments of AI than say, deploying a customer facing chatbot whereby the reputation of the company rides on every user’s interaction with a first generation technology that carries the risks of data bias and hallucinated answers. (See here for a cautionary tale.)
I almost always advise that by starting with internal processes, businesses should firstly build AI proficiency and refine their deployment strategies, before looking at customer-facing AI applications.
This ensures that when the time comes the AI, the personnel who wield it and the business processes affected are all robust enough to handle the complexities of human/AI interaction.
Balance innovation with core values
The McDonald's AI experiment should also remind us how important it is to always innovate with our core business values in mind.
As I’ve witnessed companies rush to embrace AI technologies over the past 24 months, I’ve often noticed that it's easy for them to lose sight of what truly matters to their customers and brand.
For McDonald's, speed and convenience are vitally important, but not at the expense of order accuracy and customer satisfaction.
To me, the backlash they faced in this story demonstrated the potential risks to a company's reputation by deploying AI without keeping a close focus on the core business values.
It's a pretty clear reminder that technology should enhance, not detract from, the core customer experience.
In saying this, I don’t mean to that businesses should shy away from innovation. Quite the opposite.
But, we need to ensure that AI initiatives align with and enhance your company's commitment to its customers and values.
It's about asking not just "can we?" but "should we?" with AI.
And, "how can we do this in a way that truly benefits our customers?"
I’d summarise this by saying that human oversight plays a crucial role in this balancing act. In the AI game we call this “human-in-the-loop”.
It essentially means that most of the time, AI should be a tool to augment human capabilities and core values, not replace them entirely.
By maintaining a human touch in really important business functions like customer service, businesses can ensure that their AI implementations enhance rather than detract from the customer experience.
Conclusion
The McDonald's AI drive-through experiment offers us some pretty valuable lessons as business leaders coming to grips with the first generation of AI experiments, pilot projects, and organisational deployments.
To me, the story highlights the importance of starting small, prioritising internal over external AI applications, and maintaining a hardcore focus on core business values and customer satisfaction.
If we learn these lessons, I think we can approach AI adoption in our companies more thoughtfully and strategically.
The next few years of AI deployment into the fabric of New Zealand businesses shouldn’t be about shooting for overnight transformations.
I think It should be all about careful, considered steps that create sustainable value; internally for the company, and for its customers.
To summarise what I’m saying here… embrace AI (of course!), but please do so with caution and care.
Start small, learn continuously, and don’t lose sight of what truly matters to your customers.
(I hope you’re listening Ronald?)
Got something to add? Chime in below...