Chatbots, tools to be treated with caution!

Chatbots are often seen as an easy solution to avoid overloading call centers.

However, before setting one up, you need to ask yourself three key questions to decide whether a chatbot really is the right option: at what point of the customer journey will it be positioned? Where will it be on the website and what will it look like?   


The first question is all about the risk of dehumanising the customer journey as a result of automation. You need to determine at different points of the relationship whether the user expects a human touch or if dealing with a machine is acceptable. At certain stages of the journey, all users need to talk to a real person who can show some empathy.  What’s next? You then need to carefully check there aren’t any other bots. Today, this is an issue we face: there are multiple bots at the same business… getting in each other’s way! As a result, customers might not know which bot to speak to and businesses lose coherency   


You need to examine the customer environment and technical environment in depth. People focus a lot on functional scope and the way in which the bot will manage to understand people, at the expense of its integration. However, this is a key underlying customer journey issue.  For example, if the bot opens a new tab to redirect you to another page: how do you guarantee the continuity of the relationship whilst allowing the bot to appear in this new space? As to this issue, the first question to ask yourself is not necessarily the most obvious: where should the bot appear, to be visible when the user needs it? Likewise, how should it look to make sure that people understand the help it can provide? Other, more complex questions can then be asked such as: in what circumstances should the bot redirect the user to a human operator, and to whom? Or even how do you manage customers’ understandable frustration when they’re told about a major delay or technical issue?   

Plus, the bot operates in a technical environment that it needs to understand. This is an essential element of industrialisation yet is often forgotten at the start of bot projects. Likewise, the bot’s deployment can raise confidentiality and GDPR compliance issues, particularly if personal information is provided. How will it access the data it sometimes needs to answer questions? More simply, does it meet the business’s security standards?  

The best solution to handling these issues is to properly take account of the business and technical aspects, ensure they don’t clash, and above all include a test phase proportional to the scale of the project. The creation of a test guide will allow you to make sure all the aspects of the bot are tested and educate it more easily.    


The biggest danger of a poorly-developed and poorly-integrated bot is doing damage to the brand’s image: a bad bot is a business that responds poorly to its customers. A recent Forrester survey showed that users tend to increasingly approach bots with caution… It’s not a subject to be taken lightly! In addition to the technical side, good support must provide expertise linking front and back. Frontend needs to guarantee a high-quality experience and a bot that gives appropriate answers. Backend must provide resources and an environment to easily integrate it. For businesses, chatbots are a genuine challenge in terms of reputation and public image.