Best practices

Successfully implement a chatbot with watermelon

May 11, 2022 5 min
Header Blog AI chatbot herschrijving (1)

Successfully implement a chatbot with watermelon

Leestijd: 5 minuten

We can imagine that you have many questions about the implementation of a chatbot. What does a process like this actually look like? Don’t worry, we got you and guarantee that it’s definitely going to be a success. How? With the Watermelon onboarding we ensure that you are completely prepared and know all the ins and outs about our chatbot builder. That way you won’t have any surprises. In this blog we tell you more about what you can expect during this process. Ready?

The onboarding process at Watermelon

Our goal is to have customers working independently in the chatbot builder as soon as possible. To make this possible, we have developed a special onboarding. We guide the customer through different phases so that the implementation is going to be successful.

Setting goals

We always start with a kick-off meeting. Here we determine the goals, KPIs and the scope together. There are plenty of reasons to get started with a chatbot. Time to put these on paper and think about when the chatbot will be successful for you. Of course, we’re happy to help so we’ve listed some examples of possible goals:

  • Reducing incoming emails, phone calls or chats
  • Reduce customer service costs
  • Make customer service employees work more effectively
  • Gain more insights into customer data
  • Increase customer satisfaction
  • Reduce customer turnover

Once your goals are set, you need to start making them measurable by using KPIs. Of course, you want to know whether your chatbot achieved the goal or not. Some examples of KPIs are:

  • Automation rate
  • Cost savings
  • Effectiveness of employees
  • NPS / CSAT
  • Customer turnover
  • Number of new leads

With KPIs you not only determine the success of the chatbot, but it also becomes clear where improvements are needed. Together with us you determine the scope of your chatbot. What should the chatbot be able to do and which questions are most relevant? You don’t launch everything at once, but work continuously towards a new version. This way you make it manageable and you can experiment. Do we need to make adjustments or additions? Then it will be included in the next version. We continue to discuss this with you and see where the priorities lie. Determined the goals, KPIs and scope? Good, then you can really get started with your chatbot. We will now discuss the next steps.

Blog afbeelding vragen drukte

Getting started with the platform

Time to start with the platform. First you need to log in to your Watermelon account. Are you logged in? Then the other users can be added. You are not going to handle the calls on your own right?

Have all users been added? Well done! Then you can start connecting the channels. The chatbot can be connected to the following channels:

Channels all connected? Then you’re ready to think about the personality of your chatbot.

The look and feel of your chatbot

In order to set the personality of your chatbot, there are a number of steps to follow. The look and feel of your chatbot is a crucial part of the customer experience. Think about the chatbot’s name, his or her face, the tone of voice, but also smaller details like personal interests and your chatbot’s hobbies. The personality of your chatbot should match the brand identity of your company. It’s the first impression customers experience of your service, so this needs to be a good one right away!

Adding conversations

Yes, it’s almost time to activate your chatbot. It just needs all the necessary knowledge to help your customers in the best possible way. With our easy no code chatbot builder, you can easily build your chatbot step by step. There are two ways to add knowledge to your chatbot: conversations and via FAQ. Conversations can be used to create entire dialogues that customers can have with your chatbot. Customers can start a conversation by asking a question, to which the chatbot will reply and even ask follow-up questions or other follow-up actions. The second way to add knowledge to your chatbot is by using FAQs. Where conversations are used to create complex dialogues with follow-up questions and actions, you can use FAQs to quickly answer questions using your chatbot’s AI.

During this whole process you are of course not alone, but our onboarding specialists are there to advise and guide. They plan a weekly meeting to answer questions and take a look at the building process of the conversations. They will help to make sure that the go-live of the first version goes entirely according to plan.

The Release

Once all the relevant knowledge has been added to the chatbot, it’s time to start training the chatbot. This is an important part because this is the time when your chatbot actually learns the knowledge you have added. This is a fully automated process that you just need to start. Depending on the added amount of knowledge, it takes about two to three minutes for the process to finish. So time to relax a bit.

Once all the relevant knowledge has been added to the chatbot, it’s time to start training the chatbot. After the training, it is important to do another round of testing before the actual launch. This can all be easily arranged and checked from the Watermelon platform. Everything ready to go? Then it’s time to press the big green button.

Mismatches

Building your chatbot remains an ongoing process. Especially with Conversational AI. It happens that your chatbot doesn’t recognize certain questions from customers; we call this a mismatch. These mismatches need to be analyzed. Together with you we determine whether a mismatch should be included in the current scope or added at a later stage.

What is the time investment?

First of all we have to mention that this is different for each customer, but we can give an indication. We aim to have the first version live within 30 days. In the beginning, the time investment will be a bit more. You should think of 6-8 hours per week. After the first version it’s mainly about optimizing your chatbot. Then you probably need 2 hours per week and after the first half year only 1 hour per week. The chatbot becomes a part of your work, but you get a lot in return. With this time investment, most clients are already at a 70% automation rate within the first year.

In this article we have briefly summarized what an onboarding process looks like, but there is of course much more to tell. Ask one of our colleagues for a demo so you really know everything about the implementation of your chatbot.

request demo