Sentimer (from Spain) is the Startup at the top of the ranking in the EDP Starter Acceleration Program 2017 that took place in Madrid in May. They developed a chatbot that can reduce human intervention (and costs) in several low value activities performed at front office level.
A chatbot “is an artificial intelligence (AI) program that simulates interactive human conversation by using key pre-calculated user phrases and auditory or text-based signals. Chatbots are frequently used for basic customer service and marketing systems that frequent social networking hubs and instant messaging (IM) clients. They are also often included in operating systems as intelligent virtual assistants” (techopedia.com).
In this case, the designed chatbot has been implemented in a webpage in where potential customers can make contracts related to the EDP-Carrefour offers. As we can see in the image below, the bot asks the user about their actual situation, and some options are offered in order to continue the conversation:
![](/sites/default/files/1_0.png)
![](/sites/default/files/2.png)
After some questions and answers made asking for more information from any of the parts, the conversation can end in an online contract or the user can request to be called and finish the deal with a human operator.
This solution is based in AI and Machine Learning. Before carrying on, some definitions are needed.
Artificial intelligence (AI), is “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings. The term is frequently applied to the project of developing systems endowed with the intellectual processes characteristic of humans, such as the ability to reason, discover meaning, generalize, or learn from past experience” (britannica.com).
Machine learning “is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it to learn for themselves” (expertsystem.com).
In other words, this program emulates a human being (in this case, an operator), but not only that, it also learns on its own from the data and answers received to give better results.
Machine learning techniques used by Sentimer allow to automate learning and to customize interaction with every client. There are two ways of interaction: A free conversation box and pre-defined itineraries in which the user can click directly.
The objectives of the project are:
- Incorporate the Sentimer chatbot in the EDP-Carrefour microsite.
- Validate the machine learning of the platform, performed in an autonomous way.
- Compare the results (leads) using chatbot or click to call, and with other chatbots.
- Compare this development with other chatbots developed in EDP.
This is done with the help of both EDP (Training data, customers, energy business expertise) and Sentimer (Proprietary AI for NLP (Natural Language Processing) and personalization, chatbot tech, UX design, ...).
Project results have been successfully met, with a software developed (fully operational since August 22nd) and an evolution AI tested, generating more than 10,000 conversations during the testing period, and obtaining a very positive assessment from the clients who interacted with the chatbot. Customer satisfaction is at 70 % and engagement increased.
With these results, the generated conversations seem to be useful, since in 25% of cases the user shows interest in some operational aspects (c2c, contracting process, etc.).
Once the implementation of the proof of concept has been completed, the next steps are analyzing the convenience of continuing with the test, making the latest improvements, finalizing the test or adopting the solution definitively.
This chatbot is an example of an automated and smart assistant to provide customer support and advice, improving the current situation.
These are some of the conversation examples available right now:
![](/sites/default/files/3.png)
![](/sites/default/files/4.png)