PIDAS AG Study shows: Chatbots seem to be on the Rise

insights · Veröffentlicht am 3.11.2021

Chatbots are being used more frequently. Even if cautiously. In the PIDAS AG Chatbot Benchmark Study published in 2017, the technology did not yet play a role. Today, three years later, just under 10% of the companies surveyed use chatbots.

There is still a lot of room for improvement here, that much is certain. 

However, the answers to the question «Will the importance of chatbots increase in the future» show that the signs are very positive for chatbots in marketing, sales and customer support. In fact, 72% are convinced of this. 

 

So chatbots are part of our future. Almost everyone agrees on that. PIDAS AG writes about this:

Although the chatbots used do not yet meet all expectations, the companies surveyed that have already gained experience with chatbots agree that the channel holds enormous potential and will continue to grow in importance in the future.

This is also reflected in the findings of the chatbot study conducted in 2018. Customers see great potential for the use of chatbots, especially in e-commerce and with insurance companies and banks.

However, when it comes to implementing a chatbot, opinions are no longer clear. 

Do chatbots need to be automatically intelligent? Does every chatbot need an NLP (Natural Language Processing) component? Or are simple clickbots also sufficient when it comes to the use of chatbots?

Let's take a look at the findings of the study and interpret them afterwards.

 

Chatbots vs. clickbots

One of the first parts of the study is about the type of chatbots that are used. Primarily, one can decide here between AI-based chatbots and clickbots

 

According to the study, clickbots are currently still the most common method (42%). 16% of respondents already use AI-based chatbots, i.e. intelligent bots. 

11 percent decided on a mixture of clickbots and AI bots. Clickbots are therefore still in vogue and new, AI-based technology will only slowly replace them. But why is that? 

When companies decide to use a chatbot, the primary goal is to create added value. The expectations of the users should be exceeded and the company itself wants to gain experience with the new technology.

In many cases, this is already possible with simple clickbots. For example, when it comes to generating new leads via the website or quickly guiding website visitors to the desired information, clickbots do a good job.

An example use case: The companyfitpass.ch was able to increase the number of subscription sales by 20% with a clickbot from aiaibot. Kudos, Fitpass! 

The virtual assistant points out the new subscriptions to visitors in a charming way and engages them in a conversation. Afterwards, the chatbot shows the possibilities of the subscriptions.

For such a use case, the chatbot does not necessarily need artificial intelligence, so a clickbot is quite sufficient. 

 

When do AI-based chatbots make sense?

AI-based chatbots make sense especially when it is not about comparatively simple lead generation or information forwarding, but when the users' queries enter the chatbot in an unstructured way and are also context-dependent.

An example to illustrate this: 

The chatbot of PostFinance (let's say the Swiss Postbank) receives more than 1,000 user requests daily around topics like: Account closure, account change, contact info, etc. 

Users do not want to be guided through a rule-based chatbot dialog, but enter their query directly into the input field. Thanks to an NLP component, the bot can identify most user queries, assign them to an «intent» and thus provide a suitable response. Thank you, chatbot!

As with all other technologies, it is to be expected that further major advances will be made in the field of NLP. Chatbots will benefit from this and continuously learn, make fewer mistakes and become increasingly important helpers for companies.

So it always depends on the goal of the chatbot and its area of application. For some tasks, a clickbot is sufficient; for others, artificial intelligence is necessary.

 

Combine both approaches? Does that make sense?

According to the PIDAS benchmark study, 11% of the companies surveyed relied on a combination of AI and clickbot. However, this only makes sense if there is a concrete need for the combination of both.

As a rule, these are use cases in which the chat users first enter their question using free text and the bot records it. When it comes to processing the query, however, the chatbot works with predefined dialog patterns and suggests various answer options to the user using buttons, as in the case of a clickbot.

Finally, the decision on the implementation method should only be made once the area of application or the concept of the chatbot has been defined. The chatbot canvas, which can be found here, is helpful for this.

 

In summary, chatbots are gaining strongly in importance. However, companies often seem unsure how the comparatively new technology should ideally be used. There is a simple rule of thumb here: If the areas of application are rather one-sided (e.g., the conclusion of subscriptions), a simple clickbot is sufficient. If the queries become more complicated and unstructured and the chatbot needs resources to think, then AI is necessary. 

Are you interested in the complete benchmark study? Just click here

Sophie Hundertmark is an independent consultant for strategic support and implementation of chatbots. She was one of the first Master students in Switzerland to research chatbots and is the winner of the Best Paper Award at the IWW Internet Conference 2018 in Portugal.

Want to know more about Chatbots?

Super easy! Just throw your email below and we will contact you as soon as possible with great insights.