Increasingly widespread, chatbots are a compelling tool for guiding visitors to your site. Built on a machine learning algorithm, they are able to answer ever more qualified questions. Does it make sense to use these messaging assistants on sites other than e-commerce sites? What are their advantages? How do you set up a chatbot on your website? We tell you everything!
Chatbots: how do they work?
A chatbot will always be available to answer questions from visitors to your website. Of course, these “conversational agents”, as Wikipedia calls them, are not initially able to hold an in-depth conversation, or even to answer questions that are too complicated. However, they are able to learn by themselves! How so? In practice, when configuring a chatbot you define pre-programmed response scenarios. And when the bot is not able to answer a question, a human on hand takes over. What is interesting is that the bot is able to interpret the correction that was made and therefore memorise new possible forms of response. Yes, the machine learns from past conversations between two humans, using machine learning. Be careful, though: the above definition broadly encompasses all existing forms of chatbot; in our case we will be talking exclusively about private chatbots to be configured on your own website.
An ultra-personalised experience for smoother communication
Although a conversation with a chatbot is governed by a human-machine interface, it is entirely possible for you to get your intention across, even if your request does not really “match” the library of questions and answers defined in the bot’s database. Depending on the keywords you type, the bot can infer what you want and therefore work out whether you are looking for coaching, advice, or information about a product. Among all the chatbots available, we note that Messenger’s is finding its way onto more and more websites. If you are wondering how to set up a Messenger chatbot on your website, all you need to do is download the Customer Chat plugin available on WordPress. Once this first step is done, you can configure your chatbot using the management API included in the package. If you would like more information on the subject, the official documentation provided by Facebook is available here.
Why pair Messenger with a chatbot?
There are various providers such as Botsify, Chatfuel and Onsequel that offer “turnkey” chatbots compatible with Messenger. The advantage is that it is very easy to configure: you simply create an account on one of these platforms and then link your Facebook account using the account key. Linking your chatbot to Messenger is worthwhile: it allows you to create a cross-channel conversation thanks to Messenger’s chat history. It is therefore possible to start a conversation on your website and then continue it on Facebook, or vice versa. That is an advantage:
- for the users of your chatbot (because they can see the history of their exchanges)
- but also for you and your community management!
What is more, Facebook gives you the option of creating custom advertising audiences. That is to say, you can define audiences of “warm prospects” or of customers who have had an exchange with your organisation, either through your chatbot or directly through Messenger. These audiences could be used in particular to qualify and broaden your advertising on Facebook — ideal material for good retargeting campaigns.
Bots capable of humanising the interaction
Connected to your ERP, a chatbot can have access to your stock, your products and even all the information you choose to give it. They are, quite simply, an internal search engine 2.0, with the added benefit that during a conversation they are able to steer visitors to your site towards different solutions. If, for example, you want to know a product’s dimensions, its price, or a delivery time, the conversational agent will scrape the structured data relating to your products from a table in one of your databases.
It is also very pleasant to converse with a chatbot if it has been properly configured. Indeed, it is often enough to click on alternative suggestions offered directly by the chatbot in order to request additional information. Here is an example with the Tidio chatbot.
Example of a conversation with a Tidio chatbot.
A relevant tool even for non-transactional sites
To answer the question raised in the introduction: yes, it absolutely makes sense to introduce a chatbot on your site, even if that site is not intended for selling online. Chatbots are above all a help for users, and it is best to see them in that light. Among other things, they make it possible to redirect visitors to deeper pages that are often inaccessible from the navigation menu. Moreover, if the sole purpose of your website is to present your organisation, you will surely be interested in the easier route to getting in touch that a chatbot provides. Because it is, after all, simpler to ask a question in a “live chat” than to hunt for the answer to a very specific problem in a list of frequently asked questions (FAQ). This is explained by the fact that chatbots can start a conversation effectively — for instance by asking you your favourite film genre, or whether you are more team sweet soy or team salty soy. We also note that these prompts are often presented as tags that you simply click on to navigate. On that subject, these tags also make it possible to place calls to action in a relevant way, all with the aim of supporting the user through their decision-making process (the buyer’s journey).
The key takeaway is that a chatbot will above all allow you to respond more quickly to your customers’ need for information, and therefore to optimise the user experience on your website. That happens through redirection to internal pages on your site, but not only that. A chatbot can perfectly well explain, for example: which situation a particular product or service is suited to, the benefit of one type of product compared with another, or how to choose the right one, and so on. So it is up to you to analyse the questions your visitors may have and to offer an answer to those questions with the help of your chatbot.
