Blog
AI in customer service: what McDonald’s recent troubles can teach us
This week McDonalds announced it is pulling the plug on its AI-powered trial of automated ordering at its drive-thrus, after a series of hilarious blunders were shared by customers online. These viral videos include orders containing hundreds of chicken nuggets, ice cream topped with bacon and unwanted packets of butter. But how is it that a company the size of McDonald’s, who have clearly made a significant investment in automating their service, managed to fail so spectacularly?
In my webinar earlier this week, I highlighted how important it is to consider the mistakes of the past when rolling out new capabilities of generative AI in your customer service function. Early in 2016, we saw the first wave of chatbots. Many businesses saw a shortcut to reduce headcount and save costs, leveraging chatbots as a crude attempt to replace humans. The results were poor customer experience, overloaded staff and a disappointing return on some big investments. They were doomed to failure not because of the technology, but because of how they were being used.
Nearly ten years on, and in spite of huge advances in technological capability, we’re still seeing frustrating customer experiences and poor commercial outcomes when AI is deployed in customer service settings.
What’s going on here? Is it the case that generative AI isn’t as useful as we thought? After all, if it can’t even take a McDonalds order, how good can it really be? There’s no shortage of articles suggesting that AI is, in effect, great in principle but doesn’t pay off in practice.
Our ten years of experience in applied AI has given us a distinctive perspective on this: it’s not the technology that’s the problem, it’s how it’s being deployed. Business cases for AI investment default to cutting headcount, presupposing that the goal is to replace human operatives. But the technology – as impressive as it is – will not fully replace humans.
The better question is – how can I make my human agents as productive as possible? What parts of their work can we streamline, automate or accelerate to maximise their productivity? This is about improving your ability to service your customers as much as it is about cutting costs. It can’t be achieved by simply slotting a new piece of technology into your existing processes.
To achieve real results when working with AI, you need to put your customers at the heart of your decision-making. Ten years of experience in applied AI have led us to develop a distinctive methodology for doing this. The rewards for those who harness AI strategically, focused on alleviating friction points for customers, will be huge.
Click here to watch my webinar and find out more.
