4 big myths about AI

The 4 big myths about AI in customer service

Every day, we hear something new about AI – what it is, how it’s evolving, and how it will impact us. 

There’s still so much we don’t know about this technology, but there are some myths we can already bust – at least when it comes to customer service. As time goes on, and rumors spread, it turns out that just like a bad AI chatbot, humans are pretty good at just making things up.

So, here are some of the myths we’re seeing about AI in customer service.

AI will take all customer service jobs

This is a common fear across a lot of industries, but it’s already becoming clear that unique human strengths will continue to play a major role in top-tier customer support. While AI takes on the more repetitive tasks, support teams have a chance to think deeper about their customer experience offering, and handle the more complex cases that ultimately boost customer loyalty. 

As Des Traynor, Intercom’s Co-founder and Chief Strategy Officer, said on a recent Intercom podcast: “High urgency, high drama, high emotion. Those are the types of things where humans specialize.”

“The less time spent on repetitive FAQs, the more time reps can spend on maximizing their individual and team’s success”

Not only that, but AI looks set to expand support reps’ development opportunities, and open up entirely new skills and career pathways. From bot orchestration, to knowledge base management and optimization, to product education – the less time spent on repetitive FAQs, the more time reps can spend on maximizing their individual and team’s success.

AI is overhyped

We get it, we also had a healthy dose of skepticism at the beginning, and after the crypto bubble, who wouldn’t? But now that we’re seeing AI dramatically easing the workload of support teams, its power and potential is impossible to ignore.

Fin can resolve up to 50% of questions, instantly. With half your team’s support queries taken care of, think of the opportunities you could realize in terms of optimizing your support team operations, enhancing your customer experience, and building relationships for more seamless cross-team collaboration.

“AI is hyped. It’s hyped to shit. And yet I think the hype’s real”

That’s why we believe that AI in customer service is being hyped just enough – not because of what it can do, but because of what you can do with the time it gives you back. Our Chief Product Officer, Paul Adams, has worked at the center of two huge technological advances, and is convinced AI is bigger than both.

“I worked at Google in the mobile team when we launched Android… I worked at Facebook at the height of social a few years later. AI, to me, feels way bigger, way bigger [than those] and AI is hyped. It’s hyped to shit. And yet I think the hype’s real.”

Customers aren’t ready for AI chatbots

Customers have traditionally had a rocky relationship with chatbots, often finding them awkward, unintuitive, and unhelpful. But the will to self-serve is strong, and considering the power and potential of AI chatbots, 73% of support leaders believe that customers will expect AI-assisted customer service in the next five years. 

That means that the next couple of years mark a crucial juncture for customer-chatbot relationships. Our recent user research showed that customers are already much more confident that an AI chatbot will understand their queries, and ultimately be able to help them. 

If you’ve been playing around with ChatGPT, you’ll already be expecting AI chatbots like Fin to converse in a more natural, human way than we’ve been used to in the past. It’s immediately easier for a customer to convey their point, establish an understanding with the bot, and get an answer they can understand and question further if needed. Not only that, if Fin doesn’t understand what a customer is asking, it can dive deeper, asking clarifying questions to figure out exactly how to help. 

Hallucinations make AI unsuitable for customer service

Generative AI’s trustworthiness was a major question when ChatGPT first came on the scene – and one of the biggest challenges we faced when trying to build our own AI chatbot. Despite the incredible potential of generative AI, there was no way we could risk a chatbot confidently providing plausible but incorrect information to our customers. 

“Now, companies across a range of industries are finding ways around generative AI’s biggest flaw”

However, as the weeks went on, we began building the guardrails and safety measures that form the foundation of our AI chatbot, Fin. To quote Fergal Reid, Intercom’s Senior Director of Machine Learning: “We’ve done a lot of work to use the large language model to be conversational; to use it to understand a help center article you have, but constrain it to only giving information that’s in an actual help center article that you control and that you can update and you can change and you can edit.” And it’s not just us – now, companies across a range of industries are finding ways around generative AI’s biggest flaw, and making their customers’ lives easier.

If Fin can’t answer a question based on the support content you provide, it will admit that it doesn’t know, triage the customer’s query, and seamlessly hand off to a human support rep. As a result, you never have to worry about your customer receiving a sub-par experience and leaving the chat with a little less respect for your brand. Either their query is resolved by Fin, or it’s passed to one of your trusted reps for some more TLC. 

For more on AI and the ways it’s shaking up customer service, take a look at the dedicated AI and automation section of the Intercom blog