At the start of 2016, mobile phone operator BASE Company was officially acquired by Telenet, one of Belgium's leading telecom providers. BASE has always been a young and dynamic company with a strong focus on a great user experience. Looking at the fact that 60% of consumers use chatbots to interact with businesses and the use of Facebook Messenger is popular amongst all age groups, the Telenet team contacted BotBakery, specialized in bringing together artificial intelligence, user interfaces, personalization and user experience through the creation of chatbots. Together with Ingenico, BotBakery’s remit was to set up Belgium’s first telecom recharging ‘top-up’ chatbot with an integrated payment module in Messenger to recharge prepaid cards.
BotBakery, part of Upthrust, a spin-off of The House of Marketing and the digital implementation pillar of Quanteus Group, proves with the chatbot its strength to translate promising digital technologies into practical applications.

Mobile payments in Belgium today are mainly done through webportals. Now, with the launch of BASE’s Top-up chatbot, there is a breakthrough to recharge in a chat functionality, with no need to leave the chat. Until now, BASE customers needed to go to the website or to the night shop to buy a scratch card or they had to download an application to recharge their prepaid card. As of today, those customers can recharge with only a few clicks within their Facebook Messenger. The only thing they need to do is to accept the privacy statement, give the phone number they want to top up, define the amount they want to recharge and fill in their payment details.

Bridging the knowing-doing gap

BotBakery built up significant experience and expertise, allowing us to develop a strong methodology to create the most effective chatbots. As a result and to share our expertise, we created our BotCanvas. To ensure the success of our BASE chatbot project, we started of course with filling out our own BotCanvas template.

The BotConvas is divided into several focus areas: customer centricity, system integrations, functionalities and design, conversation channels, performance and organizations. 

Customer centricity focus:

It’s very important to focus on the customer problem you are trying to solve and to approach this problem in a customer centric way. Several customer needs pop up when someone wants to recharge its prepaid card.

• Customers want the process to be easy, fast, frictionless and accessible wherever they are. They don’t want to download yet another application, nor do they want to run through a long process or pop out to the night shop to buy a scratch card.
• Customers want to get appropriate help if needed, via FAQ or through customer care help.
• Customers want to buy immediately, without extra unnecessary detours. They just want to recharge their card on the platform they are already on, which is possible in BASE’s Top-up chatbot now where customers top up in less than a minute.

System integrations:

We prepared the required technical integrations to ensure every functionality of the chatbot works. In this case, two internal and two external integrations were needed. We integrated BASE’s back-end systems to recognize phone numbers and to avoid a number has been entered in the wrong format for example. The customer care services of BASE were also integrated in our BotBakery platform to make sure that a smooth handover could take place if the customer needs human help.

Alongside these internal systems, we integrated Ingenico, required to transact secure payments inside Facebook Messenger, and Dialogflow, a Natural Language Processing system enabling the chatbot to provide answers to frequently asked questions.

Design & functionalities:

As every brand has its own personality and target audience, a chatbot must talk with an appropriate tone-of-voice. A chatbot can talk in a funny or serious way, a casual or formal way, an enthusiastic or no-nonsense way, can use a lot of emoticons,…

The functionalities are very versatile, but the most important ones are: the e-commerce function, the possibility to answer the main FAQs and the very smooth handover to customer care if required.

Conversation channels:

Our BotBakery platform is compatible with a lot of channels such as Facebook Messenger, WhatsApp, Web Chat, Voice Bots and SMS. Because of the widespread use of Messenger by their target audience, BASE decided to integrate the chatbot into Facebook Messenger. This decision was largely strategic, because your conversation channel should be a platform with a lot of users, but mostly it should be a platform that is widely used by your (potential) customers.

Performance:

It’s all about the numbers, right? So we designed a custom-made dashboard with the main KPIs, which enables BASE to get more information about the number of Facebook users, the number of (un)successful payments, total payment amounts and average payment amounts. We even included a report to track from which sources the users came.

Our custom-designed analytics will enable us to continue to train the bot even after the official launch. Our virtual chatbot friend will learn from its mistakes and gain an increasingly more accurate understanding of users’ problems and questions over time.

Organization:

Every integration of a new tool in a business model will have an impact on the organization and the employees. We adopted an agile approach with all the stakeholders. Through weekly Skype meetings and clear briefings we were able to make sure everybody was on the same page. This was paramount, as a lot of internal and external systems and processes were integrated, which needs alignment in order to connect all the dots. It is important to always have the buy-in of your employees by involving them in the process.

With the help of the BotCanvas, we were able to develop an advanced demo bot. This bot was launched in a closed test environment, which enabled us to run a user experience review. The chatbot was given out to be tested by a string of different customer types. Their input, the input of BASE and our own test rounds enabled us to optimize the chatbot. This optimized chatbot became the official BASE Top-up chatbot.

Leaving our fingerprint behind

With this chatbot project we left a twofold fingerprint. We helped BASE launch Belgium’s first telecom top-up chatbot with an integrated payment module in Messenger. We enabled BASE’s customers to top up their prepaid cards on their favorite communication channel without any friction. 

Hard to believe? Take a look:

Insights and learnings from the project

1. Make sure the different systems to be integrated (internal and external systems) are aligned and the right mutual links are established. If not, you’ll have a lot of unconnected dots and the chatbot functionalities won’t work. A good technical analysis is key!
2. Operate on the platform that is used by your customers.
3. If you are planning to develop a Facebook Messenger chatbot, bear in mind you’ll need permission from Facebook. This Facebook Review can take a long time.
4. Be aware that each business has its own tone-of-voice. A chatbot should definitely speak with the right one.
5. If you’re at a point in the project where you’re too immersed in the topic, you risk losing your external unbiased view. Don’t assume that because you understand everything the bot is saying, everybody else does too.
6. Finally, don’t forget to have fun! Chatbots are an exciting new way to communicate with your customers!
   
   Interested in your own chatbot? Get started with our easy-to-use BotCanvas, a template that helps you to identify opportunities and pitfalls.