Customer service automation in the DISCOVERY group – chatbots for Player.pl and TVN24 GO
55%
cases solved independently by the chatbot
> 4 000
messages exchanged daily
4 minutes
average session duration
Challenges
- Reduce the workload of customer service staff
- Provide greater flexibility to handle spikes in the number of contacts
Five years ago, when Player Customer Service launched Facebook Messenger as a new channel to engage with users, the idea seemed pretty unusual. At the time, social media were considered space for entertainment and promotional content. But the channel was to become the primary user service center for the TVN Group’s VOD platforms. Viewers have their phones on them at all times and interactions are often seamlessly interlaced on multiple screens. If they experience a difficulty while watching a film from an online VOD rental, they look for a swift solution. So they ask questions via IM and receive support at any time of the day. Thus, it was a natural development path in the customer service area. To make this communication channel work effectively, automation processes needed to be implemented.
Starting the work from the client’s perspective
The bots development began with a list of the most common questions and common issues reported by users of VOD platforms. This was later used to prepare scenarios for conversations and assistance, and thousands of messages adjusted to many variables
We started our preparations by looking at solutions to automate user service in the US and Europe. It is worth mentioning here BlueBot, a well-known airline bot, which serves as a great example of how natural and fluent conversation can be. It provides excellent linguistic capabilities and quickly guides you through the processes related to your journey. This and other examples have sharply whetted our appetite for innovation and raised expectations as regards the design of intelligent solutions, says Grzegorz Przywara, Customer Care Director, player.pl.
We had been long looking for a supplier in our market who would meet these needs. We finally came across the KODA Bots team. We engaged in many long, sincere discussions. We already knew that we had to start with a simple concept. And then start building more complex and independent solutions.
Test implementation
Julia was the first chatbot to engage in conversations with users. Together with the KODA Bots team, the client decided to test the solution on a lower-traffic channel, namely the TVN24 GO service. The bot performed well enough that they decided to implement automated communication for ten times more traffic. Radek began working as a virtual assistant in the Player service in February 2020.
The bots operated under the watchful eye of account managers. Conversations were read, and any ambiguities served as useful tips to help optimize the content.
Automation results
1. Reduce the workload of customer service staff
More than 50 percent of all reported issues are solved by Julia and Radek independently. The staff can now deal with more complex matters, but they can also join the conversation at any time and provide help to solve the problem. You can see everything in the dashboard, where messages from multiple sources gather: FB Messenger, the email inbox, the hotline, Instagram and from mobile app shops. The integration with Salesforce CRM has made the bots part of the company’s customer service system.”
2. Provide greater flexibility to handle spikes in the number of contacts
Customer service can quickly offload traffic that arrives rapidly on Friday nights. For example, this was the case with Jan Komasa’s film “Suicide Room: Hater.” The sales launch of this long-awaited title, released on the VOD service shortly after its cinema premiere, resulted in a tenfold increase in the number of user requests. At times things happen that cannot be predicted – for example, a software update in mobile devices that causes technical issues. When users report a problem, it is enough to add a few lines describing an issue, and the bot will almost immediately be ready to respond and provide assistance.
User ratings
The ultimate success factor is the users’ satisfaction with this form of contact. They appreciate the speed of assistance and its availability at any time. Once they are satisfied, they come back for more, because they know they can count on effective help. Already half of all queries to Player customer service comes from the FB Messenger application.
This figure will continue to grow. It is more and more convenient to use this channel, sometimes without even having to write, since a lot of information on the most popular topics is available in scrollable galleries. The automation platform is continuously adapting to users’ preferences.
Platform development
Future plans include broadening the pool of communication tools. Getting in touch with customer service could well be the opening of a longer relationship with the brand. Users can then receive personalized messages that exactly match their interests. Sports fans will receive a notification if they are expecting a sports event to be broadcast. TV series lovers will be notified with the date and time of the latest episodes’ release.
The opportunities are plentiful, depending on what information a user provides about their preferences. Such personalization becomes possible only with advanced tools capable of collecting such large amounts of data and holds the future of user experience building.
New opportunities of interacting
Implementing chatbots for Player and TVN24 GO shows that Messenger is a good choice for carrying out tasks in the customer service and customer care fields. It is not only the ease of managing the automation platform and the convenience of instant messaging conversations that speaks for itself. This also opens up new possibilities for building relationships with users.
To date, bots have been associated with automated, mass communication. But this technology is developing rapidly, giving us the chance to learn about our callers’ preferences and use new marketing automation formats. We can address the callers by their name, interact with them exactly when they need assistance, at the right moment in the buying process, inspire them and suggest news that will interest them. We have entered the era of hyper-personalization, and a well-designed chatbot will ensure that customers receive personalized service.