With the multiplication of communication channels and the democratization of artificial intelligence technologies, bots have become a new means of communication with prospects and customers, but also a tool to improve the productivity of customer service …

One of the starting points was the opening of Messenging solutions, in particular Facebook Messenger which authorized direct communication with its Facebook contacts, as if we were sending an emailing.

In fact, as soon as a user gives you permission to chat with him (eg conversation with a Facebook page, an exchange via a messenger chat widget on your site, a Facebook Ads advertisement, etc.), you can resume contact with him just like you would by email, with a much higher read rate thanks to notifications, the scarcity of messages…

Beyond that, messenging (chat, Facebook, Twitter, etc.) is increasingly becoming one of the preferred communication channels for young and new generations:

So if the traditional channels (ex: email, phone…) always remain essential for important interactions (ex: big technical problem, litigation…), for small simple requests, consumers prefer to have an immediate response to their requests and not wait any longer.

But that’s not all, a chatbot is also a tool that improves customer service productivity.

According to a study published on The echoes, bots could save 27 billion hours by 2023:

Therefore, if used well, the chatbot can become an excellent tool for automating recurring customer service tasks. The key is to identify in customer requests requests that can be automated, that have a high volume (mini 100 / day) and that require a little personalization.

Below is a simple example that shows the potential savings that can be made in a customer service department that manages 1000 requests per day (regardless of the channel), with 30% of requests that could be managed by a bot:

Be careful, however, to think that the Chatbot will replace your current tools and your customer service agents … and magically answer all the questions of your customers and prospects.

Thee self-learning cat bot does not exist… you must first show him how to solve the problem, which involves initial training (understand a design of the conversations, scenarios and scripts), and integration with your information system.

Then do not think that the chatbot will solve your internal dysfunctions … on the contrary, it could accentuate them! Indeed, if you do not currently know how to manage a request manually (eg: incorrect information in the information system, complex responses to data with a share of creativity or thought …), the bot will not be able to do better. On the contrary, it could generate frustrations due to irrelevant answers.

To bring enough added value, you must find one or more use cases (the use cases) which generate volume, and which require a personalized response if possible.

The chatbot is just one more tool in your arsenal to improve the Customer Experience, with a new communication channel or a new way of solving a problem.

So sometimes a simple FAQ can be more than enough if it is a question of giving generic answers.

The graph below shows the different advantages and disadvantages of the different customer service channels:

2 practical cases of using a chatbot to improve the customer experience!

The starting point for setting up a chatbot should be to identify a customer problem to be resolved.

This problem can be:

  • Curation, that is, to solve a problem or respond to a request in the most efficient way possible;
  • Prevention, that is to say anticipating problems via onboaring, coaching …

Here is an example of one of our customers at Eloquant.com who works in the energy field.

Its recurring problem is being able to handle a huge influx of requests during cold periods (December to February), with peaks of requests from one week to the next depending on the cold waves.

During sudden cold waves, customer support has enormous difficulty in dealing with short peaks of customer requests in a short timeframe, hence the establishment of a bot which is capable of taking gas orders. in self-service, to do meter reading …

The bot thus complements the customer service team to manage simple tasks, with a large volume, while being personalized.

The bot also offers personalized services with a personalized context that would have been difficult to offer otherwise.

Thus, at one of our other customers, Méteo England, we warn of the time of day every morning at a fixed time via a Facebook Messenger notification.

The benefit is simple: offer via the best channel (the smartphone) contextual information (the city weather) at the right time (just before getting dressed).

This function is very often used by parents who want to bring their children to school, and who want to know before how to dress them, or simply to know how to dress before going to work …

20 questions to ask yourself before launching your chatbot!

The establishment of a chatbot can have a real impact on the efficiency of your teams, it is essential to start by asking the right questions …

Here are 20 questions to ask yourself in order to frame your chatbot project in your customer service:

  1. Have you defined the purpose of the chatbot, and the need to solve (curative or preventive)?
  2. Does this problem have a minimum volume / day, with personalization if possible?
  3. Did you frame the project, with a budget, a project manager …?
  4. Have you estimated a possible ROI through the automation of certain tasks taken into account by the bot, or by the best conversion / retention of your customers?
  5. What are the 3 major KPIs that will make the project a success (eg: the number of bot conversations, the satisfaction score, the percentage of understanding, the decrease in the number of emails received by the after-sales service, etc.)?
  6. Do your competitors already offer a bot (on Facebook, on their website…)?
  7. Who is the persona who will use the bot, what is the usage scenario and why should he use the Bot?
  8. What is the perimeter of the bot (what it can do and what it can not do, the language …)?
  9. What are the 3 to 4 uses cases with a minimum of added value and volume that you have defined, and will the chatbot be able to answer them (ex: connection to CRM or ERP …)?
  10. What are the chains of questions or the conversation necessary to get the answer to the question?
  11. What will be the personality of the bot (your…) and its design?
  12. Where will the bot be placed (Facebook page, website, etc.)?
  13. What are the triggers for the visibility of the bot (according to pages, according to behaviors…)?
  14. What is the bot’s communication channel (text, Messenger, Whatsapp, Google Home, Alexa, Skype…)?
  15. What is the impact on the information system (eg, reading, writing, etc.)?
  16. How is the chatbot going to integrate the current system (eg escalation to a human in the event of misunderstanding…)?
  17. Who will be the users of the Chatbot, and who will make it evolve (the bot trainer)?
  18. Have you planned a post conversation bot evaluation, and what will be the scenarios?
  19. What is the internal (to reassure the teams) and external (to ensure adoption by customers) communication mechanism?
  20. Have you planned an internal and external beta test program?
  21. What dashboards will you define and to whom will they be sent?
  22. How will you develop the knowledge of the bot, and what are the other problems that you can entrust to it in phase 2?

Do you want to go further?

Feel free to leave me a message in the comments to tell me about your use cases (Or on Eloquant.com) in order to see if on the one hand your situation requires the use of a Chatbot, and then make an estimate of the ROI of its implementation.