Customer service in the future: how important is the human role still?
Vasco employees Joep van der Velden and Jeroen Severs talked to Leontien Sterk, Director Customer Operations at CANAL+. About customer satisfaction and the future of customer service. What will be the human role in it?
Leontien, how do you know Vasco?
I worked with your CEO, Ruben Uppelschoten, for many years. At least ten. That was at UPC at the time, which later became VodafoneZiggo. After that, Ruben set up Vasco Consult very successfully. I did some other things first and started working at CANAL+ a few years ago. Vasco also did some assignments here, when it was still called M7. In 2019, M7 was acquired by CANAL+, but because of covid, they did not actively start rebranding right away. They have since started doing so, with the Netherlands being one of the first countries.
What does your job entail?
I am responsible for Customer Operations. I deal with everything to do with customer contact: customer care, call centers, chats, the self-service part. Not just for the Netherlands, but for all European brands under CANAL+. In recent years, I have also started working more with colleagues from the entire CANAL+ Group, from other parts of the world. With those colleagues, I work a lot on digitalization, AI implementation, measuring quality over large volumes.
This rebranding to CANAL+ that has now been done for a number of brands, has that had any impact on customer service or operations in general?
Not too bad, I notice that everyone is positive about it. It is always a chore to do though, but if you prepare it well and involve everyone in time, it is not too difficult. It is important to have the list ready with what all needs to be changed and that determines how easily you run through it.
And those are trajectories you manage?
For customer care, I do it together with the local managers. I am not their direct manager. That has to do with local legislation. But for customer care, they report to me. It really does feel like my team. Including intern and product manager, there are seven of us.
And if you translate that to operations and customer care, how big is it?
The local teams together are about 17 people. Some teams like the Czech Republic are a bit bigger, because that's a big market. But I also have someone who does Hungary and Romania on his own. He manages these two countries from Romania. In this call center, there are agents who speak both languages. So, it's a small team. And everything else is outsourced to external partners.
What channels do you use for customer service?
Phone, chat, e-mail. We are pushing very much on chat. This varies by country, depending on how some countries, cultures deal with it. Chat does start to become more and more mainstream in terms of communication.
And is that live chat or via a chatbot?
Both chatbot and live chat. We try to give the chatbot as much knowledge as possible at the back end, so that it can capture as much interaction as possible. That piece of knowledge management is a challenge, though. And on a global level, we are working on AI. Between the chatbot and live chat, we are trying to stick AI. To automate even more.
We have a proposition for you: ‘The success of a customer service team depends heavily on the level of autonomy employees experience.’
Yes, totally agree! No patronizing call center agents. The more you can solve at the front end, the less customers will be sent from pillar to post. I am an advocate that with self-service and one-time human contact, as much as possible is solved for the customer directly. So, autonomy has to do with that. Agents don't have to pass on to someone who knows better or is allowed more. For example, crediting. In some organizations, the agent must hand over everything above the amount of a tenner to a supervisor. I like a flat organization, where as much as possible can be solved directly.
What is not complex should not be made too complex?
Exactly. I actually quite dislike the classic call centers. Of course, that sounds very strange when you are responsible for them, but I try to prevent customers from ending up in a call center. That's why I try to keep them as small as possible. What you see with classic call centers is that their business model consists of the number of hours they sell. The more hours taken, the more they earn. On the contrary, I want to move towards less contact with the customer. I don't want that customer to call me. Only if he wants to buy something. It's always a hassle when the customer calls you. Then something hasn't gone right. I am looking for call centers that are trying to get that volume down. For example, how can we ensure that the chatbot gives more answers than our staff?
What does that mean for the agent's work?
By removing the different levels with a chatbot and AI, you are left with slightly fewer employees. But they are allowed and able to do a lot. This allows you to help the customer directly in one go. In addition, agents can update the information they collect themselves in the knowledge base and train the bots. This is also how I want to manage teams; I put the autonomy and problem-solving with the teams themselves. I try to create a culture where everyone feels free to give feedback to each other.
Another statement: ‘In a successful customer service organization, customer satisfaction is more important than strictly following business procedures.’
I find this tricky; it depends very much on the context. For 80%, customer satisfaction is more important. But it could be that you work in a financial organization, where you are really bound by rules. Or it involves privacy-sensitive information. Data privacy can be very limiting. If someone wants to convert something or is moving house, you might want to arrange something quickly with the data. But that’s restricted.
And how is that at CANAL+?
With us, rules are often exceptions. You try to build a lot of procedures around the customer and make him as comfortable as possible. The dilemmas are often about data privacy. Or the customer wants something that just really can't be done. Sometimes it's about the limitations in the system. Or the conditions don't allow something.
Another statement: ‘Efficiency should always be a priority over customer focus to ensure scalability.
To ensure scalability, I don't so much agree with that. I think efficiency and customer satisfaction can go hand in hand. Inefficient processes are often, in my view, not the most customer-friendly processes either. Suppose a customer disagrees with an invoice. And it takes a very long time for that to be sorted out due to inefficient processes. That's not very customer-friendly because the agent can't resolve it right away.
For you, what are the key KPIs and metrics to monitor customer service performance?
We measure customer satisfaction, the NPS is slightly different. Every time a customer contacts us, we measure satisfaction. The NPS is measured less frequently and is more about the overall recommendation. The ‘first-time fix’, this one is very much about autonomy. Being able to solve everything right away for customers, so they don't call back. And we also measure the ‘self-service percentage’. To what extent are customers able to solve problems themselves. For example, we measure the percentage that cannot figure it out with the chatbot. Those customers are redirected to the live chat. The percentage that then gets stuck goes on to calling an agent. We try to measure the whole process. We have almost completed that. Then I can really see how good the information delivery is to customers, so they don't have to call us anymore.
So, you really operate at that tactical and strategic level to work with local managers to see how things can be improved?
That's right, we are positively surprised at how much the chatbot can already solve. How many questions do customers ask the chatbot, what kind of questions and why do they ask those questions? In turn, we need to store and share those questions somewhere else, so that in the future the customer doesn't even have to ask the chatbot. That way, you're working every time to make sure customers just stay away.
How do you ensure that process optimization has a positive impact on the customer journey?
We call this the call prevention program. That is actually process optimization. I learnt this from Ruben at the time and when Vasco did an assignment with us. Vasco focused on ‘the first 90 days of the customer’. We then ask ourselves: what is that customer calling about? And we improve our processes on that. We go through the whole journey. What communication does the customer get, what kind of letters or e-mails are sent. If we know what the customer is calling about, we can make sure he doesn't call again. The triggers are often a letter, an e-mail, an invoice. Or in the beginning, when a customer signs up, setting the password.
Do you look at the differences between your companies; for example, does the chatbot score better at one than another?
The Netherlands is digitally very mature. We have good internet, so we are in a luxury position here. In countries like Hungary, things are different. People there were calling every month about their bills. Then we thought: why do you call every month, because that invoice is the same every month. They often paid it late as well. For that market, we now send a text message prior to the invoice. We can't send people to a chatbot if they hardly use the internet. So, we send them to an IVR (interactive voice response - a telephone choice menu). There, we help them via a cassette tape with instructions about their bill. That way, we can help those customers more easily and efficiently. That is a totally different approach than in the Netherlands.
Do you think these cultural differences will stay that way for the time being?
Developments are going very fast now. But in some countries, you still have disadvantaged areas. You have villages in Romania where one person in the village has a mobile phone and he arranges everything for the whole village. Local call centers know their market and customers very well and respond appropriately. They understand the culture. In the Netherlands, we say to the customer: go and chat instead of calling. We can't do that in Romania.
A final statement: ‘AI and automation can guarantee customer satisfaction better than human interaction can.’
Not now, but my guess is in 3 years' time.
And where do you see the biggest development to take place now?
Right now, it's mainly in training the chatbots and AI bots. Feeding them with knowledge. Making sure the bot does not derail and stays within the legal frameworks. You need people to do that. Everyone thinks that if you have AI, you can say goodbye to humans and save time and money. But in my experience, that is not the case.
At your company, is there a big challenge in capturing people's knowledge to feed AI?
At CANAL+ almost all knowledge is captured in knowledge bases. The challenge lies mainly in the fact that we have knowledge in different languages, that it is secured differently per country. We are also dealing with very new technology. It can do a lot, such as generating content. But to ensure that an AI bot can help the customer in 99% of cases and within all legally safe frameworks, we are not that far yet. There really is still work in that and we need people to do that.
Are you already working on adding these people to the team?
Yes, we are working on that. Profiles of people are also changing because of these developments. The agent who used to maintain the knowledge base now has to train the bot as well. If a new product comes on the market or there is a new campaign, we do need to know that that bot is giving the right answers. And that is not so well developed at the moment.
So, do you see a growth in teams dedicated to this very kind of development?
Not yet, because that has to do with costs. But I do see a development in the profile, both in call centers and in management positions. It becomes part of your job responsibilities. AI takes an awful lot of time to test. We are now live in two countries. In the Czech Republic and Slovakia, it was quite a challenge. There, the AI bot finds it difficult to tell those two languages apart. Keeping certain technical documentation apart, AI also struggles. Because the tool has so much trouble with it, we decided to take out technical documents for now. While that is what you want to achieve, that the tool can answer all questions.
Is the AI solution you just described a further developed chatbot? And what does it do more than a chatbot?
We have both tools. The ‘old-fashioned’ chatbot uses machine learning. Then our people must programme the bot completely for what it needs to do. The AI bot is much more dynamic and less pre-programmed. But we must give it clear frameworks. That it will only search a certain type of document.
How exactly is the AI tool applied at your company? Does it answer customers directly or does it support employees in giving the right answer to customers?
We have now set it up as an intermediate solution. The customers ask their questions directly to the chatbot. And if that fails, they go to the AI tool. We do this to test it in a safe way first before exposing customers to it.
So, is there any human verification in between before AI feeds it back to the customer?
The human, the employee, intervenes when the AI tool starts giving errors or hesitates too long. Then there will be a live chat with the customer.
You mentioned at the beginning that you aim for the customer to call as little as possible. What does that customer himself think about that? We sometimes hear on assignments that the customer would like to speak to someone directly.
In our customer satisfaction, I don't see that reflected. We have a very good customer-satisfaction score. And as long as that score is good, I am satisfied. We also notice that customers' behavior is changing, and they are getting more and more used to not calling. The biggest irritation among customers is when it takes too long for them to be helped or get an answer. Then you quickly get a bad customer experience.
Looking back on our conversation. For you, what are the most important ingredients for successful customer service?
The people. Whether you do self-service or take a lot of calls, in the end it's the people who do it. Whether you have a small, flexible team - which I think is ideal - or a large team; in the end, the people are most important to me. I also try to visit the call centers as much as possible. Then I am not in the meeting room, but just among the people. Listening to what's going on and sensing the atmosphere.
One last question. Will external call centers always exist?
I think the chances are slim. Unless they evolve with the new technologies, becoming agencies that handle chat and social media for you. But if you can digitize a lot, in principle you no longer need classic call centers with hundreds of people answering the phone. So that becomes more like a small team monitoring everything and making sure all channels are functioning properly.
Leontien Sterk has been working at CANAL+ since 2022. Before that, she worked for other companies in the telecom and media sector for many years.
