2017 was a year in which the keyword “chatbot” became more and more a topic and many trade media intensively dealt with the development and also the acceptance of chatbots. Especially in the social media channels this topic has spread rapidly. We as a shopware agency would now like to take a closer look at whether these chatbots are useful within the shopware world.
Let’s start with the basic question of what a chatbot is. Put simply, a chatbot is a dialogue-based communication system with a computer system (chat – bot = chat robot). Means that the user enters a text in a chat window and the computer responds to it.
For a long time, since the first development in 1966 (ELIZA), these chatbots were based on a rule-based system and the input (also called intention) had to correspond exactly to the specification stored in the computer, so that the predefined answer could be played out. If you entered an unknown or incorrect question, the system was not able to answer it. 99.9% of the 100,000 chatbots on Facebook Messenger are already rule-based.
Only the development of programmed artificial intelligence has significantly changed this circumstance. With regard to chatbots, this approach of artificial intelligence is also called “Natural Language Processing” (NLU). In this way the bot automatically learns to understand an intention in very different formulations in order to give the correct answer. An autodidactic system with amazing fault tolerance. In contrast to the old system, this takes a lot of frustration away from the user, who can thus reach their goal more quickly. This is exactly where we would like to emphasize the benefits for e-commerce, especially for shopware shop operators.
What you could use a chatbot in Shopware for
Every Shopware Shop sits on a real data treasure, which remains unused in most cases. A chatbot is active 24/7 in a shop and could have access to the entire database. These would be e.g. all articles with all deposited information or evenly also all orders of a certain customer. A chatbot could be integrated into the shop like a classic chat widget.
Statistically, up to 35% of all inquiries within a shop are identical and could be answered automatically. As a result, significant quantities of inquiries are regularly received that bind “human resources”. With a chatbot, exactly this share could be processed automatically around the clock.
A customer would like to know how much shipping to Austria costs. The chatbot can now access the shipping cost configuration of the Shopware shop and determine the exact price. Within a few moments the customer receives the exactly fitting answer to his intention without anyone having to become active in the company of the shop.
One of the most important arguments of the local retail trade is personal product advice. For a good service it is important to help the customer with questions and recommendations. This creates trust and binds the customer. If this personal consultation is not available or is not carried out well or too late, this can lead to frustration of the customer in the retail trade.
In online retailing, there is now a natural distance to the customer. The chatbot will not be able to completely replace a competent employee, who stands in front of the customer, on the emotional level. However, they can score around the clock with short reaction times, comprehensive product knowledge and spontaneous learning ability.
So he can also become proactive and address the customer directly (e.g. Good evening Mr. Müller! How can I help you today?). This binds the customer to an interaction, who can now enter his or her intention, such as “”Hello. I need new black sneakers in size 48”. For example, the chatbot reads the following in this intention: “”Sneaker” = category sports shoes, “black” = filter group color – value “black” and “size 48” = variant group size – “48”. It then displays a list of all articles that match these values. In other words, a recommendation from the chatbot.
The extension “Shopware Product Streams” built into a chatbot would look like this: Suppose you are selling baggage and would like to provide your customers with a list of suitcases that are subject to Easy Jet’s hand baggage size guidelines. You create a product stream based on these dimensions. The customer gives now the intention “I fly next week with Easy Jet to Berlin and need a suitcase, which goes through as hand luggage”. The chatbot now uses your “Shopware Product Stream” and plays the items in it in a separate listing.
Since Shopware version 5.3, Shopware provides a very powerful combination of the “Customer Streams” and “Advanced Promotion Suite” extensions.
The chatbot can also approach the customers here and refer to current promotions. Customers, in turn, can ask for active offers on their own initiative. For example, if we have a promotion for the category “Sports Shoes” and one for the Customer Stream “Sports Shoes Regular Customers”, the chatbot will notify those customers who match the Customer Stream of this promotion.
Since the end customer in the Shopware version 5.3 always remain logged in, you can be addressed by the chatbot at any time personally with their names.
A chatbot also for Facebook
As early as 2016, Facebook opened its own Facebook Messenger for third parties, clearing the way for chatbots. 1.2 billion people worldwide currently use Messenger, a relatively gigantic number. Now imagine a customer can interact with your shop without being in your shop!
Facebook Broadcasting aka Newsletter 2.0
Every shop operator knows the challenges of email marketing: delivery times, individual content, white label servers, spam, too many newsletters, etc…
All this can improve with a chatbot. If a customer from your Shopware Shop is linked to your Facebook Messenger Chatbot (DOI), you can also send proactive messages from Shopware to Facebook Messenger. There are currently no spam filters here. Customers no longer need to actively open their e-mail inbox and browse through several e-mails. As soon as the message has been sent automatically and, above all, individually, the message pops up as a push notification on the customer’s smartphone. Here, too, the powerful features of the Shopware customer streams can of course be used and thus your customers can be segmented according to their target groups.
Alexa and Google Assistant = chatbot
Last Christmas, the best-selling item on Amazon was the Echo Dot. In contrast, Google announced that since October every second a Smart Speaker (Google Home and Google Home Mini) had been sold. It is currently assumed that in Germany alone at least 1 million households have at least 1 Echo device. The spread of smart speakers is therefore much more advanced than many would have thought.
At the recently finished CES in Las Vegas, it also became clear where the journey was heading: refrigerators, headphones, microwaves, cars, ovens, etc. will be equipped with either Alexa or the Google Assistant. In some places, the manufacturers even install both assistants in one device. It will therefore not be long before digital language assistants are introduced in almost every household. After the advent of e-commerce in the early 2000s, the mobile wave since 2007, the “Conversional Commerce” is probably approaching us inexorably.
Alexa Skills and Actions for Google
Both systems offer the possibility to extend the respective digital assistant by extensions of third party providers. Alexa calls these extensions “Skills”, Google calls them “Actions for Google”. If you have already integrated a chatbot in Shopware, the way is (technologically seen) not far anymore to offer a chatbot on the digital language assistant.
Imagine the following scene: A young father is standing in the kitchen preparing food and following the instructions of a recipe skill through his digital language assistant integrated in the fridge. His hands are dirty from cooking. His little son comes into the kitchen and shows him his broken football shoes and points out the important game next weekend. Now it has to be quick and the young father immediately remembers the shop where he ordered these shoes a while ago.
Now he addresses the digital assistant in his refrigerator:
“”Hey Google, ask “Kids-Sportshop” which soccer shoes I ordered 2 months ago”. Thanks to the direct connection to the shopware shop of “Kinder-Sportshop” the chatbot knows immediately who the request is, which order and above all which product it is. “”Hi Max, you ordered the model 2 months ago “Super soccer shoes” for 49€”. Max: “”Please buy again”. Chatbot: “”All right, the model “Super soccer shoes” will be shipped immediately”.
This is how we see the not too distant future with regard to customer interactions in e-commerce.
We are currently at a turning point in e-commerce. Due to technological progress and above all the advance of large providers (e.g. Google, Facebook and Amazon), digital language assistants and thus also chat bots are becoming more and more “socially acceptable”. However, it must also be said that this technology is still in its infancy. Now and in the foreseeable future, one must not make the mistake of suggesting to a customer that one would interact with a real person here. This may not be possible until 5-10 years from now.
However, as digital language assistants become more widespread, users are made aware of the need to speak and write with artificial intelligence. After all, time and comfort is an increasingly important factor in all our interactions. Therefore, in the near future it will probably no longer be sufficient to operate a shop exclusively in the form of a “website”.
Pixup Chatbot for Shopware
The article I have written here is of course not entirely selfless. I have to admit that I was confronted with the term “chatbot” almost daily from January to May 2017, but I didn’t care any further. But then my curiosity became too great and I got involved in the topic. Appropriate was that in June 2017 the “Chatbot” Summit” took place in Berlin and I had a close look at it. Already at this fair the decision was made to bring something like this into the shopware world.
Since then we have been working with our own development team on the integration of a chatbot system for shopware. All the above examples will be made available here in Q1 2018 as a plugin to the entire Shopware community via the Shopware Community Store. Of course not only rule-based, but on a NLP algorithm and thus on a self-learning system. The deep integration into shopware and above all the ease of use for every shop operator, which offers the respective end customer a clear added value, was of great importance to us. The entire system is designed so that each shop operator can tailor the chatbot to his or her target group and needs. Third party plugins can also be integrated, so that for example the “Store Locator” of Netinventors can be used within the chatbot.
In the coming weeks we will go into more and more detail about the individual features and possibilities here in the blog. Be sure to stay tuned.