How intelligent automation can empower the customer service chatbot

Photo by Ali Abdul Rahman
author

Chatbots can be fun, but they can route customer service queries to the wrong department too. If they were only more intelligent…

I contacted Apple Support yesterday. Unable to match my query with the standard categories, I was eventually led to a text chat. I believe Apple failed the Turing test for the initial greetings. I’m certain it was a bot because later on during the chat, the messages changed and more grammatical mistakes were revealed. Unless we're now adding poor English to our chatbots, I believe I was eventually in conversation with a human. But I shouldn't be too unfair. Representatives of Apple Support are remarkably chipper and put you in such a good mood, you often forget what the original problem was. I've found myself playing the same game, matching their chirpiness word for word, virtual hug for hug.

If it was indeed a chatbot the whole time, I was fooled. The benefits of an automated customer service process, should you get them right, are many. Cost savings is one of them. While Apple may not worry so much about its bank balance, other companies do. By reducing reliance on real people to handle thousands of calls, you can save big. For instance, McKinsey reckons you can save up to 40% on service costs by revamping the customer experience using technology, and that’s not small change.

What I’m not advocating is completely automated customer service. You still need a human touch when and where it matters. And the potential for automation to improve the customer journey in this respect is huge. Imagine chatbots handling most customer queries, reducing the turnaround time of solving simple issues, while making sure that complicated queries get through to a real person. Speed and efficiency gains overnight! Yet, there’s more automation can do than simply follow rules-based criteria.

Efficiency gains according to the customer

As business owners or employees, we can be quite myopic. Our problems are all that matter, so when we look to increase efficiency for instance, we do so with as much detail as we can. The most important thing is more often than not what the business can gain, and customers come second. What a shame! What a missed opportunity! If you flip this around and focus on efficiency gains for your customer, you stand to benefit too. If you have a customer service function, this is the chance to revolutionise the way you interact and build relationships, and change the way your customer sees you.

One of life’s shared bugbears is waiting on hold. Or in a queue. Or not being able to contact an organisation at all. Good implementation of RPA has done a lot to remedy how long we wait for customer service solutions, and there’s more to do. Here’s a scenario you may not have thought of: support ticket triage. Rules-based automation can assist in routing support tickets so that they reach the right customer service department, but what if the issue is more complicated? What if the query is routed to the wrong help desk?

One of the bottlenecks of customer service support ticket efficiency is having to re-route incorrectly tagged queries. And it’s usually a manual process. Intelligent automation, on the other hand, offers a perfect solution, and in real-time.

Let’s say a customer is talking to a chatbot and the bot is trying its best to make sure the query reaches the right team. For example, “My iPhone has a cracked screen”. A simple rules-based process may route the query to the iPhone sales team. A more intelligent automation process will quickly find and add labels and categories based on previous knowledge, assigning tags to ensure the customer query is taken care of by the right person. And in super-quick time! It will assess the language used, such as “cracked screen”, and because of previous knowledge, it knows to route the customer to the Apple Care team, or perhaps continue the conversation by asking how long they’ve had the device. Should the customer be outside the warranty or Apple Care period, the bot recommends screen replacement services without ever having to route the query to a real person. And the whole issue is tagged correctly for reference.

The sooner you can satisfactorily resolve a customer query, the faster you can move to the next one. And that customer should feel not only satisfied that you helped to resolve their issue, but that you did it quickly. And everyone had reason to be chipper.

Back to blog list

Fast track to machine learning starts with a free Aito Sandbox.

Address

Aito Intelligence Oy

c/o Innovation Home

Toinen Linja 14

00530 Helsinki

Finland

VAT ID FI28756352

See map

Contact

COVID-19 has driven us all to work remote, please connect with us online. Stay safe & play with data!

About usContact usJoin our Slack workspace

Follow us