The power of AI is undeniable, but can they replace call centres?


When we hear the words AI and call centres, we immediately imagine dialling a number, only to be greeted by a semirobotic voice (think Siri or Google). While it’s true that AI can be more efficient than humans (at some things), does it really mean that human call centre agents are becoming obsolete?

I don’t think so. Personally, when I call into a call centre, I much prefer to speak with a human being than endlessly pressing numbers at an AI’s instruction. There’s something about human interaction that can never be fully captured by an AI.

In our digital age, call centres remain the front line for many organisations. It’s the channel customers turn to for the most complex or urgent problems. Call centre agents also humanise the organisation, in good ways as well as bad. A positive experience with an individual agent equates to a positive experience with the organisation. Likewise, after a bad experience, we are probably going to say something like “x company is terrible."

So how can we ensure we provide the best human experience possible? The answer is with AI, but not in the sense we mentioned earlier.

Sentiment analysis

Voice to text technology can record what the customer has said much faster with high accuracy than a call centre agent manually taking notes. Sentiment analysis can be performed on this data to identify if a customer is getting increasingly frustrated or angry.

Can you imagine the difference between angrily asking to speak to the manager vs. the agent proactively suggesting, “Let me get my manager on this so we can sort this out for you as soon as possible”?

Capture data from customer interactions

I, like many customers, get so frustrated when I am redirected to another department and must repeat everything I said to the previous agent.

This could be easily prevented by using the data we recorded previously to monitor sentiment and using AI to identify the nature of the call. Text analytics – in particular, natural language processing (NLP), can generate the type of call (complaints, query) and the topics (account, billing). With this information in hand, the next agent will be better prepared for the call.

Forecast call volumes

What’s the longest time you’ve been placed on hold waiting to speak to an agent? Personally, my frustration grows with every passing second while listening to the (annoying) on-hold music. By the time I get to an agent, I am already angry and impatient.

Unexpected call volumes can cause havoc in a call centre. Forecasting techniques can use historical and real-time data to make predictions about incoming call volumes in the next hours, days or weeks. With this information in hand, call centres can improve their workforce planning, which means no more long hold times.

Agent matching

We, as individuals, are good at different things, and call centre agents are no different. Agents have different skills, and they also connect with different types of people. We can use machine learning to identify the strengths in each call centre agent and match the caller to the agent who will provide the best customer experience.

The power of AI is undeniable, but can they replace call centres? Read Kelly's post on the answer. Click To Tweet

Let’s say I contact a call centre. I have a complaint regarding my billing. The AI can look at my record and see that I had a great interaction with call centre agent Jane Doe in the past. I rated her 10 out of 10 in a post-call survey. Jane isn’t available right now – perhaps she’s moved on. AI checks Jane’s profile, and she has a bubbly demeanor. It can then search the available agents and find similar personalities, and filter for those who are great at solving billing issues. I have another great experience and get my problem solved. That same bubbly personality that I connected with could have annoyed another more laid-back customer.

When AI is used to support the call centre, instead of replacing it, we are creating better human connections. Humans are the real power behind every advancement, and they can enhance their potential with the help of AI.

That’s what we mean when we say that “together” the possibilities are exponential.


About Author

Kelly Lu

Kelly is a specialist in the Advanced Analytics and AI practice at SAS South Africa. Prior to joining SAS, she has worked in Analytics in the Retail, Banking and Insurance industries. With this experience, she advises businesses on best practices and what is required for a successful deployment of analytics. She is particularly enthusiastic about helping businesses find the unusual use cases, that makes an impact to their customers.

Leave A Reply

Back to Top