What Is The Reality Of The Customer Service AI Revolution?

Although artificial intelligence (AI) has been explored for many decades it was the recent release of ChatGPT by OpenAI that really caught the public imagination. Finally, most people could see and easily experiment with AI in a way that had not been possible before.

The famous computer scientist Alan Turing asked ‘can machines think?’ back in 1950. He created a test, which became known as the Turing test, to determine if an AI could pretend it is a human by having a conversation with a human without them realizing they are talking to a bot.

This is how many customer service systems are designed today. You get service immediately 24/7 because a bot answers the call or text message. The bot attempts to help out and will pass the customer to a human if the conversation becomes too complex.

The implication for customer service is that this new wave of generative AI - which has only really emerged this year - will finally be good enough for customers to have long conversations with a bot using natural language. The message is clear - most customer service interactions could be automated and therefore fewer human agents are needed.

Some people are making this argument. Jon Stokes, the co-founder of tech news site Ars Technica, said: “It’s a really easy yes,” when asked if AI will replace customer service jobs.

But this view is something of an outlier.

McKinsey suggests that the real benefit AI brings to the table is automation and productivity. Customers want 24/7 instant service so give it to them with better and better conversational bots, but always hand off to a human if the customer interaction is not working well.

This view emphasizes the ability to do more with the same budget - answer more customers, answer them faster, and keep humans focused on the complex problems that AI struggles with. This is all positive because it means the customer service workers are not all being replaced by robots, but it also means that customers are getting a better experience. The human agents also get to handle more complex (and interesting) problems.

A recent article in the Harvard Business Review goes even further when it states:

“Despite the media narrative to the contrary, generative AI will not wipe out entire categories of jobs, such as those in customer service. Automation is ideally about unlocking human potential to do tasks differently and do different, higher-value tasks.”

I think the real point here is that AI will replace some tasks or processes, but not necessarily jobs.

A contact center agent needs to write up notes and categorize calls after talking to a customer - get the AI to do that. Supervisors need to monitor for difficult encounters so they can maintain quality - get the AI to listen to everything and alert the supervisor to what needs checking. Agents need coaching to improve weaker skills - get the AI to advise and guide what needs to be improved. Use AI for specific tasks where it can act as an assistant.

There are dozens of processes like this where AI can immediately be used as an assistant to human agents. It not only boosts productivity, but it helps the agents to be better trained, more confident, more supported and reduces the mundane repetitive processes in their work.

This is where the generative AI change will be immediate.

Customer-facing chatbots will follow soon - and will quickly improve - but there will always be new problems the bot has never seen or problems that the customer struggles to explain.

The reality for AI is a mix of what McKinsey and the HBR have suggested. We will see customers served faster and more capable bots that can handle more and more routine problems, but we will not see a complete replacement of human agents. Humans are still better at managing human problems than any AI system.

Author Katrin Langley is the Vice President of Business Development with Awesome CX based in Tampa, Florida.