Eleven Hundred Agency

  • LinkedIn LinkedIn
  • Facebook Facebook
  • X X

We Talk Tech

Eleven Hundred Agency

Our AI got it wrong

Navigating the promise and peril of AI in customer experience

By Harry Mottram

I recently had an unnerving moment when following up with a retailer's customer service team. After getting in touch about receiving the wrong order, I was assured that the mix up had taken place due to the company’s “AI getting it wrong”.

Points for honesty, of course, but it did strike me as a bit strange. It wasn't just an apology for a mistake; it was an acknowledgment that the very technology meant to streamline and improve customer service was, in this case, the source of the problem.

This made me realise how deeply AI has woven itself into our everyday experiences, even in scenarios that require the human touch.

Perhaps that’s exactly where we currently are with the state of AI in customer experience: brimming with potential, yet still capable of hiccups. With AI becoming increasingly integrated into our daily lives, how do we embrace its strengths while keeping a critical eye on its limitations?

AI's potential for better service

It’s no secret that AI is shaking up the world of customer service. From chatbots and virtual assistants to recommendation engines, AI is changing how businesses talk to their customers. According to Gartner, conversational AI is expected to manage about 25 percent of all customer service interactions by 2027, compared to only two percent in 2022. That’s a huge leap, and it shows how much companies are banking on AI to deliver better service.

The benefits are hard to ignore. Customers can get help any time of day, responses come quickly, and businesses can tackle customer queries more efficiently. The efficiency gains for businesses are immense, as are the opportunities to improve customer satisfaction by resolving queries more effectively.

When AI drops the ball

However, like my experience of receiving the wrong order, AI isn’t immune to mistakes. Machine learning models are only as good as the data they are trained on. Biases, errors, and gaps in training data can inevitably lead to mistakes that frustrate customers.

Concerningly, my interaction with the customer service assistant, who was helping resolve the issue, illustrated the tendency to blame technology when something goes wrong, rather than taking accountability for its decisions. When implemented carelessly, AI can inadvertently create a divide between businesses and their customers, ironically undermining the very experience it aims to improve.

When that happens, it’s not just frustrating, it can chip away at trust.

Getting AI right

The key here is balance. AI can’t replace the human touch—and it shouldn’t try to. The best customer service happens when AI and human agents work together. AI can take care of repetitive tasks, like pulling up customer history or tracking an order, so that humans can focus on the things that really matter: empathy, problem-solving, and building relationships.

The phrase “our AI got it wrong” isn’t just a quirky excuse— for me, it’s a sign of how much things are changing. AI is no longer some futuristic idea; it’s part of our everyday lives. For businesses, the challenge is figuring out how to use AI in ways that make life easier for customers without losing sight of what makes service truly great: a human connection.

For comms professionals working with tech brands, this is a big moment. It’s a chance to help companies talk openly about what AI can and can’t do. Customers value honesty, and they’ll respect businesses that are upfront with the strengths and weaknesses of their AI. Finding the sweet spot between innovation and trust - and getting that right - will shape the future of customer service in this AI-driven world.