About Melanie Müller and the Benchmark Study

Melanie Müller is Head of Marketing at Pidas AG, the Customer Care Company. Pidas is a service company in the customer service sector.
Among other things, she co-supervised a benchmark study series that Pidas has been conducting for several years together with the Zurich University of Applied Sciences (ZHAW). This series of studies is about customer service in the digital age. This year, the benchmark study focuses on customer orientation, service automation and chatbots. To this end, 210 companies in the DACH region were surveyed.

What are the most important findings of the study?

In the 2018 study, they looked at chatbots from the user’s perspective. This year, they wanted to find out what entrepreneurs think about chatbots.
The Bechmark Study 2018 found that users are willing to communicate with companies via chatbot and see the benefits of chatbots.
They were therefore surprised at how few companies use chatbots. Only 10% of the companies surveyed currently use chatbots.

Why do too few companies use a chatbot?

One of Pidas’ theories for this is that companies are unaware of the benefits and added value of chatbots. These advantages vary from use case to use case. In the service case, these include
● Availability 24/7
● Users can handle their service requests 24/7 in self-service
● Great customer experience
● The company’s workload is reduced when standard questions are dealt with directly by the chatbot
Another theory is that there are false expectations towards chatbots. The topic of AI and robots is often addressed in the media, and in some cases the image has emerged that AI is already very close to our human abilities. This raises expectations of the chatbot. This can also be seen in the figures. Of the companies surveyed that use chatbots, 60% were satisfied. This means that 40% of the expectations of the chatbot were not met or the use case was too broad.

What is the solution for more companies to use bots?

Companies simply don’t know enough about what a chatbot can do and what benefits it brings. That’s why a lot of educational work still needs to be done.
When you sit down with companies, you can show them how easy it is to integrate a bot and how quickly you can derive added value from it. A good step is to start with a pilot chatbot that is live quickly and with which you can gain initial experience. It also helps to put together an interdisciplinary team as early as possible and involve it in the development process.
A chatbot is not static, but lives from being developed further. This is why companies must not stop with the first version of the chatbot. It is therefore important to support companies so that the chatbot is a success.

Should chatbots replace live chats and what does this have to do with the first-time resolution rate?

The first-time fix rate is an important key figure. It indicates whether a customer issue was resolved the first time.
As part of the study, they picked out the live chat and the chatbot and compared the initial resolution rates of the two. The first solution rate of the chatbot was just under 50%, that of the live chat at 30%.
A chatbot is particularly good when it comes to one topic. The broader the topic, the more difficult and complicated it is to solve with the chatbot. A good solution for companies that already use a live chat is to place the chatbot in front of the live chat. The chatbot can then ask the most frequently asked questions and possibly solve the problem. Only then does the transfer to the agent take place at the user’s request. Gradually, you may then be able to cover all questions with the chatbot.

Click bot vs. AI bot – what’s the point?

The study showed that click bots are the most common. It was exciting that almost a third of the interviewees were unable to give an assessment because they had perhaps never asked themselves the question: is this just a click bot or does it have AI components?
It is important to first consider what exactly you need the chatbot for and only then look at whether a click bot is enough or whether you need one with AI. In most cases, a click bot is sufficient. A chatbot with AI components is more effort. For a good, AI-based bot, you need a lot of structured training data so that the chatbot has a benefit for the customer and doesn’t say “I don’t understand you” every second question. You probably need 200 data records for each topic that the chatbot is supposed to answer automatically.

Where are the most exciting fields for chatbots in the future?

In the study, 70% of respondents say that they see a future for chatbots and that their importance will continue to grow. This is also how they see it at Pidas.
The topic of chatbots should be viewed more broadly. The chatbot is the first step. If it works well, it can be networked with third-party systems, for example. You can then also think about automation, the chatbot could store data in other systems, query order statuses, etc.
The field is really broad, but you should always consider beforehand what you really need, what is feasible and what really adds value.

You can read more about the PIDAS study in the following blog post.[vc_empty_space height=”40px”]

[vc_empty_space height=”40px”]And if you don’t want to miss any more podcast episodes, then sign up for the mailing list at the bottom of this page.