How AI and Machine Learning Can Reduce Contact Centre Costs

Tony Dolan of Connect Managed shares some examples of how contact centres can save money by introducing artificial intelligence (AI).

The concepts of artificial intelligence and machine learning have been around for some time, but we are just beginning to see their power to transform the way we do business.

These technologies can be used to improve many of your business’s processes and functions, but their effect is most visible in the world of customer service. Specifically, when properly applied, they can help lower the costs associated with running a contact centre while also helping you deliver an improved customer experience.

To give you an idea as to why this is the case, consider the following ways AI and machine learning can be integrated into your contact centre’s processes.

Customer Experience

We all know providing an excellent customer experience is key to pleasing people and winning their loyalty, and customer service is in a unique position to deliver on this important objective, largely because it has so much direct contact with customers.

AI and machine learning services, such as Amazon Connect, can help you do this (but only if you have the right partner to implement it in a right way).

Consider the following example:

You’re at the Airport…

…and your flight has been cancelled.

An airline advisor tells all passengers not to queue at the help desk but to call customer services, who are aware of the cancellation, to reschedule your flight.

When you call, instead of having to navigate through an IVR menu and explain the problem to at least one agent, an automated agent (such as Amazon Lex) says:

Hi! We see your flight has been cancelled. The next available flight departs tomorrow at 9am. Would you like me to book you a seat?

From the contact centre perspective, the result of this is twofold: first, the customer gets their problem solved almost instantly and without further frustration; and second, you solve that problem without having to involve an agent, which saves you time and money.

Further steps could be added to delight your customers too. For example, for loyal customers you could instruct Lex to offer airport vouchers for the inconvenience or a discount on their flight.

Without AI, the customer would have had a much more frustrating experience, and this could cause them to vow never to do business with you again.

But because their problem was solved quickly, they are more likely to use you again and tell friends and family about the great customer service they received in an otherwise frustrating experience.

The Technology Behind AI and Machine Learning

Amazon Connect offers some technology including Amazon Transcribe and Amazon Comprehend.

Transcribe allows you to capture valuable data between your customers and agents in a text format rather than voice for quicker and better analysis. For example, identify recurring keywords and topics to update self-service tools such as your FAQs page.

Comprehend is a natural language processing (NLP) service that uses machine learning to find insights and relationships in text. For example, run sentiment analysis to judge which customers are happy with the service they received and which were not.

In this sense, AI and machine learning should be viewed as tools to empower your agents to do better. Many people see these technologies as replacements for people, but this simply isn’t true.

Nothing beats the human touch, but when utilising AI and machine learning, your agents will be better equipped to provide service levels that can enhance your customer experience.

More Efficient Outbound Calls

Customer service is not always thought of as a driver of sales. But this is changing, especially now that customer experience has become so important.

AI and machine learning can be used to help you do this. For example, let’s say you’re working for a bank and you learn there is a high dropout rate in online applications.

AI can be configured to send an email to those who abandon the application asking them if there was an issue or if they need help completing the process.

This additional contact shows customers you care, and your assistance could be just enough to push them to complete the application or at least gain a better understanding why customers are not completing the application.

Furthermore, because the process is fully automated, no additional resources are needed. In the past, you could employ a strategy such as the one mentioned above, but you would need people to independently reach out to those who had abandoned an application. This not only takes more time, but it also consumes your resources, often making the effort not even worth the hassle.

However, with the AI and machine learning, you can identify issues and address them more quickly and efficiently than ever before.

Marginal Gains = Increased Efficiency and Reduced Costs

The concept of marginal gains has revolutionised some sports. But the same approach can be used in business too.

The doctrine of marginal gains is all about how small incremental improvements in any process can lead to a significant improvement when they are all added together.

We’ve hinted at this a little, but now it’s time to dig a bit deeper into how AI and machine learning can help you reduce the costs of running your contact centre without sacrificing the service you provide.

Without AI

To illustrate this, let’s consider the following customer enquiry from the credit card industry.

• A customer has reached their credit limit.
• They have made a payment to clear some of the balance so they can make more purchases.
• However the system takes 2-3 days to process a payment, so the customer won’t be able to use the card until the payment clears.
• Unaware of this, they call the contact centre to ask why they can’t use their card.
• The agent makes them aware of the 2-3 day clearing period and advises them to check their balance before they try to use the card again.
• End of the call.

Although this is a very simple and quick query for the agent to deal with, what happens if you have 10,000 calls like this every month? Even if each call takes 30 seconds, it would account for 83 hours of agents’ time to deal with every month.

In the theme of marginal gains, what if you have multiple examples of this type of query? If you automated all of them, that would add up to a serious gain in terms of time and cost.

With AI

Now let’s look at the same scenario with AI technology in place.

• As before, the customer calls the contact centre to ask why they can’t use their card.
• AI sees the customer’s balance is close to their limit so asks a qualifying question, “I can see you recently made a payment on your card. Does your query relate to that?”
• The customer replies “Yes” so the automated agent continues: “OK. This payment will take 2-3 days to clear. Do you want me to send you a text when it does?”
• The customer replies “Yes”.
• End of the call.

The customer receives the same information but the call never reaches an agent, hence saving your resources for more complex calls. The text alert is an example of how to delight the customer as well as ensure they don’t call back with the same query to ask for an update.

Escalation to a live agent is always an option, but for simple customer enquiries like this, automation needs to be considered.

It’s important to remember that any customer enquiry that doesn’t reach an agent but has a positive outcome will save time and money while also pleasing customers.

Conclusion

For AI and machine learning in customer service it is very much the art of the possible. We have illustrated a few examples here, but the possibilities are truly endless.

Each business will be unique with its challenge so the same technology will be used differently, but the business outcome remains the same, increased efficiency and improved customer experience.

For AI and machine learning to have a real impact on your customer service operation, it needs to be properly implemented.

You need consultancy and expertise to help you fully leverage the power of these technologies and layer them with your existing infrastructure and processes. Only then can they improve the efficiency your business.

Author: Robyn Coppell