The emergence of chatbots with conversational intelligence

Chatbots are now “conversational intelligences.” Just like their virtual neighbors, voicebots, callbots or humanoid robots. They represent one of the areas of expertise of artificial intelligence. Their number of creations is growing and their sophistication continues to be refined. This makes it possible to give today high-performance virtual assistants.

The burgeoning chatbots, voicebots and humanoid robots

According to the CNIL study (1), “artificial intelligence is the scientific field that deals with the study, design and implementation of ‘smart machines’.”

Artificial intelligence is a form of data automation to help and relieve human activities. It is a continuation of the IT revolution. The latter was responsible for creating many jobs. Indeed, understanding and optimizing computing and Internet development to mobilize important resources.

Artificial intelligence is experiencing the same development. Just after Great Britain (2) raised 581 million euros of investments in artificial intelligence startups in 2016, France had raised 278 million euros (2). Placing second on its continental scale, France was already realising its commitment and interest in artificial intelligence.

Thus chatbots, voicebots and other tools with conversational intelligence were gradually born.

Their overall work process is relatively simple and similar for all. These virtual intelligences were all created with the aim of solving problems quickly through the same process. The first step is to automate the cognitive task, which is to associate the perception of a thing with a situation. For example, watching a traffic light go green is associated with crossing the road. By dint of numerous repeated associations, a database is then created. An algorithm is then made. It splits previously collected data into two parts so that it can gradually slide into a learning process. This is when the machine is subjected to tests that will allow to organize the data and judge its relevance.

Performance, if nothing

After all these steps, the machine must achieve a performance of more than 98% if it wants to provide real help to humans. Aloes onlymen, the AI will be directed towards data collection. This will allow it to adapt and provide more and more precise answers to its users (3).

Some industries are heavily involved in the development of artificial intelligence and the evolution of chatbots. These include the transport sector and the banking sector (2).

The impact in the transport sector is mainly related to vehicle automation. The project of a fully autonomous car, i.e. in charge of both the execution of the steering and the supervision of the external environment, is being carried out. It is one of the most high-profile and converged artificial intelligence missions.

As far as banking is concerned, customer relations are at the heart of the issues. The automation and comfort that artificial intelligence could bring to customers are two missions in the middle of a project.

In Bordeaux, an expert in artificial intelligence revolutionizes bots

With the collaboration of Ai3, a Bordeaux consulting and integration company specializing in digital transformation, Dydu has developed a chatbot of a new generation. He created the first sculpin that allows employees to manage their Microsoft work environment.

Its role is to simplify people’s working lives. For example, help with the use of Microsoft suite tools such as Outlook, Excel or Planner. Zen solves these recurring problems. With its database regularly updated by Ai3, Zen helps focus on complex tasks.

Dydu is behind many artificial intelligence innovations such as the voicebot that answers the user’s questions vocally, or the domobot capable of controlling objects connected by voice.

It is now possible to create your own bot with Dydu. These bots allow you to quickly create robot swith conversational intelligence. Multiple uses: customer service, visitor reception, transactions, home automation… and many others.

Towards self-learning AI

Algorithmic technologies have evolved considerably since their beginnings around 1950. As the needs and techniques have undergot enormous change, technologies are also adapting to our operation. At first, the techniques reproduced human intelligence via artificial intelligence. Since 1980, humans have been gradually trying to create algorithmic techniques that allow machines to learn directly: “machine learning”.

Since 2010, the “deep learning” consists of allowing self-learning of machines (3). As a result, AI is beginning to outperform the best human experts in some areas. A transformation that turns to the conversational strategy of these machines, an emerging market.


Dydu release

  1. Why France needs a CNIL digital culture of July 2017
  2. Artificial Intelligence and Work of France Strategy for March 2018
  3. Artificial intelligence, the state of art and perspectives for France, Min. economics and finance, Min. cohesion of the territories, 2018

Abonnez-vous à la Newsletter