Customer service success revealed by Elisa

This is a compressed excerpt uncovering the most informative questions from our interview with Mailiis Ploomann, Head of Telecom Services at Elisa Estonia. Read about the kinds of obstacles Elisa has overcome on the way to becoming a customer success innovator. We also touch upon the most efficient ways to solve accompanying issues, and bring out how to reach top results in the industry.

Mailiis started her journey with Elisa 10 years ago as a Product Manager. Several years later she became the Head of Telecom Services. One of the greatest projects she has led was customer service automation through AI technologies which made her an innovation leader in the field.

If you’re from a large company with lots of customers, you probably know that poor customer service kills business. The most dangerous situation that can happen in that case is when it becomes impossible to keep relations with current customers, and there’s no point to even talk about acquiring new ones. After purchasing a product, customer service becomes the client’s key point of reference. It’s the actual communication with the company that builds customer experience.

Mailiis, what is the most important thing in customer service at Elisa?

The most important thing for any telecom company is that you would be there for your customer at the exact moment when your customer needs a solution, whatever question or problem they have. But it’s not that simple. There are other components in this equation: to do it cost-efficiently, sustainably. This means it’s good to depend less on human labor. Not because there’s something wrong with it but because in the future you won’t be able to meet customer demand with it. And that’s the main reason why we’ve chosen to automate customer service.

How many customer contacts does Elisa have per month?

It varies a bit on the month of the year. But it would be 100,000 to 150,000, if we count all incoming contacts in all channels.

The first innovation you introduced to your customers was the chatbot software, Annika. Does a company need to make any preparations for innovations with AI?

You need to answer a few strategic questions before starting to automate anything: why do you need it, who is responsible for it, who’s the most interested in the fact that you have customer service, and who benefits the most internally? So ‘the customer’ is not the answer there. And when you answer all those questions, most probably you’ll end up with a decision to make. Because continuing to do the same old things in the same old way and expecting a different result – that’s just nonsense. You have to do things differently internally, and decide who’s going to be the one making sure that whatever procedure you’re automating actually works. And then you must line your company against that strategic decision.

You mentioned that in the beginning 10% of customer service was automated, what’s the current situation?

Today the bot handles over 30%, but not yet 40%, of all the incoming contacts. It recognizes a lot more, around 75-76% of all the contacts but we’re not ready to implement those dialogue tree flows just yet. And out of all inbound calls and all chats, if we consider all the incoming written contacts to be chats, 60% are actually handled by the bot. If we talk about all the incoming contacts in general, I think it’s 40% of all the inbounds.

How is Annika different from other chatbots?

When we think about Annika today, we don’t talk about a chatbot anymore. I think that’s the main difference. It’s automating all our customer service so there are different AI products put together, different models and applications working together in order to create a smooth customer journey. This is Annika today. But even when we started with a chatbot, initially we already created it a bit differently than all the other chatbots out there. I suspect that most chatbots are built just to answer a particular question. Its aim is to recognize the question and to give some sort of answer, just to get rid of your problem. That has never been the aim for Annika or any of our chatbots. Our goal has always been to automate the customer journey and fully resolve the customer contact. So Annika never gives a general answer or a link ‘go there’, it always takes into account the whole journey you have to go through for your questions to be fully resolved. So the approach is different. We concentrate on seeking for a solution, never just an answer.

Why do you think some companies do not see business value in customer service automation with AI?

There are few successors out there and it’s so easy to choose the wrong track. Let’s take a chatbot example. It’s so easy to create a plain chatbot that does nothing for your business. Customers are not happy with that, so you conclude that the chatbots don’t work. It’s not the chatbot that doesn’t work – it’s your wrong approach. You have to rethink the way you approach this technology. It shouldn’t be seen as IT or as a program. Yes, all those things are required, but first you have to rethink the way you do business. And for that, you need inspiration, success stories, someone who says it’s okay. I think AI technologies are at the beginning of the maturity curve, there are companies that do it, half of them or even fewer succeed. Those are the ones to watch out for and learn from because they’re building the actual future and the future that is worth copying. It’s smart to copy the companies that are succeeding. 

Why did you choose MindTitan?

I think trust is always built between people. And in a way, the trust was there or it appeared very quickly because people from MindTitan were the first ones to listen to what sort of problems we had. They didn’t try to sell us the product that they had stored and thought about it as the coolest thing ever because it’s AI. They didn’t need us to adapt to their product, it was the other way around. They listened to what our actual problems were and then built a solution for our actual business problems. That’s the key to most of the issues we have. And most of the projects are so successful thanks to that.