The hype about chatbots does not abate. At the latest since Alexa, Siri and Google Assistant turn on the lights in our apartment and read out or read news, not only such voice-based bots are socially acceptable. Even messenger services such as Facebook Messenger, WhatsApp or the good, old SMS become more important for the customer approach and more and more brands use here Chatbots. But why? And what should be considered when planning a chat bot?
Chatbots are used in many cases in customer service and where customer centering has become the central element for brands, fixed “consultation times” just are no longer up to date. If I have a problem with my internet connection at 11 pm, but the call center closed three hours ago, that is rather limited customer service. By the same token, not every business can afford to offer a 24-hour service, and this is exactly where chatbots play their first strength.
Chatbots do not sleep
Whether integrated on their own website or in a messenger service, a chatbot is available for inquiries at any time of the day or night, rendering the well-known “The next employee available to them in 42 minutes” obsolete with matching queue earwigs. He can react directly and help with the problem. And here the second factor comes into play, because most service inquiries are often trivial in all sectors.
Appointments, information about services, the billing status or the question of why a red light on the router has been blinking for hours are classic examples. Such requests account for much of the service center’s capacity, both on the phone and in stores.
With such simple things, chatbots can be a real relief and offer a WIN-WIN: The customer gets the solution to his problem at his preferred time, without any waiting time.
The customer service and therefore also the brand are perceived as competent and fast (and not least innovative) and also from an economic point of view make such electronic chatter lines sense. In addition, trained employees can then be used for the “real” problems that the chatbot passes on if, in exchange with the customer, they do not know how to help them.
Automated product consultation by bots
Especially in Asia, for example, the mother of all apps WeChat is already very far. There, customers can request a variety of information about products via chatbot and then order them right away with the messenger and pay for it via the integrated system.
They have taken on the needs of customers and provides them in the tough competition with other brands with chatbots a comfortable service. An example, that one can look in the often called “service desert” Germany. Because not only the Asians who are traditionally very open to any technology want to define their type of brand communication.
Especially those targeting Digital Natives, Millennials, Generation Z and other spongy generic terms may not want to call when they have a problem. Or run into business if there is a problem with a contract. And finally, we all have that desire for more service, less waiting time, and lower-friction processes.
In fact, chatbots can be a perfect complement to customer communications and, by the way, not just for service requests, but also, as in WeChat’s Asian example, can be used to recommend individual products on request, such as a style advisor bot.
For example, he asks in chat what the occasion is, asks for clothes size and things like “conspicuous” or “elegant” and then can put together an outfit, including a link to the suitably filled shopping cart. The fashion giant H & M does the same in a similar way and eBay also offers a product search service.
In addition to news, weather and recipe Bots- the ingenuity here for other industries basically virtually no limits. Above all, considering how self-learning AI’s will expand such chatbots in the future and are already doing so.
How do I build a chatbot?
The technical aspect is left aside, there are a number of things to consider in advance in planning and conceptual design, so that a chatbot can offer real added value for existing and potential customers. Here are the most important things:
1. Define target groups
The definition of target groups in the form of personas is the starting point for planning a chat bot. It is important to determine who the bot is supposed to talk to, how these target groups are communicating, and what will probably be their pain points or requests. The creation of personas, as it is done in other marketing disciplines, is therefore the starting point.
2. Define tasks and measure performance
A chatbot can not and should not be everything. He does not have to give restaurant recommendations in the immediate vicinity and at the same time tips for choosing the right engine oil. He should cover an (extended) task, be it the answering of service requests or the recommendation of sneakers and not much else, but then well thought out. Here, the bot should respond purposefully and not start thousands of random attempts.
Chatbots associated with customer accounts allow the customer’s preferences to be tailored to personalized requests. And to offer him a real added value and not a nice gimmick that can write text in the Facebook Messenger.
How successful the chatbot is should then also be measured, either in the form of feedback from the users or in the form of hard KPIs. For example, when and how many conversations led to a sale.
This is the only way the bot can be run and ultimately improve the business and increase customer experience through better personalization.
The definition of tasks sounds trivial, but it pays off to set clear tasks and play through possible question-answer combinations. A look from the outside is just as important as there are likely to be a lot of things written into the chat that you did not think about. So that should be intercepted as well.
3. Charmed Chatbot
The fact that a chatbot should not pretend to be human is one thing, but that he does not face the client as a faceless pile of source code is the other. So with the above points, you should decide which personality of the bot is most likely to appeal to the target audience and if its nature fits in with its tasks. To miss the chatbot of a face or a body in the form of a (3D) model, a cartoon figurine or the like, is generally perceived as more sympathetic. The same applies to the choice of name and the way he writes.
And as with us humans, not only does that make the difference, but also with what “story” the chatbot accompanies this part of the customer journey. What does he say when and what does he say when the customer has asked for something specific? How does he react if he does not understand the request at all and when does he know it might be time to talk about certain content? For example, buy recommendations and advice, as well as human employees in sales must be able to. The personality is an important component that can decide on the perception of the bot and thus of the brand.
4. Testing, testing, testing
You do not have to be a computer scientist to know that you should not throw untested products on the market. While minor errors or absolutely ineffective features are often noticed only in productive use, the logic of the chatbots should be checked beforehand. Once the target group, the tasks and the answers as well as the character have been determined, it makes sense to have the bot tested outside the filter bubble of the planning team.
On the one hand, whether the answers make sense and the Chatbot the user sensibly accompanied. On the other hand, it also depends on whether he looks like “from a piece”. So is he talking to the right target groups, is his personality consistently authentic, and is the chatbot doing the job in line with his character? Often you only notice these things when the results of the planning come together and then adjust.
With these four cornerstones, a framework for a chatbot can be developed to build an effective and profitable communication tool. And as is the case with people, one should also teach him more tricks in the long run, always with a view to the wishes of the target groups.