Artificial intelligence: the solution to transform your data into value

After the Internet, artificial intelligence technologies are generating the fourth industrial revolution. Economic models and information systems are being turned upside down. Information, and the knowledge that derives from it, is today the key differentiating factor between companies. Meeting with Guillaume Bréjaud, General Manager of oppScience in charge of the development of the bee4sense activity.

But how do we define artificial intelligence? “In reality, when we talk about artificial intelligence, we’re talking about augmented intelligence, because we’re not yet in a world where the machine is going to take power over man. This augmented intelligence consists of mathematical means called algorithms, allowing organizations and individuals to process very large volumes of data”. However, it is the Internet that directly created the large volume of data to be processed and the problems involved in processing it. The volume of information has literally exploded and its processing by simple rules has become impossible. This is where AI finds its application and addresses all sectors of activity. “The real break lies in the fact that the algorithms can provide results that are sometimes not very comprehensible, which puts the user before the obligation to accept what the machine says.

 

The challenge for companies today is therefore to move from a model where software governs data by copying its constraints, to a model in which data governs software. This means rethinking the architecture of information systems. “Information comes from different origins, takes different forms – text, data, images, videos – and is stored or distributed in different spaces such as databases, on the web, within applications. Even before using the information, the system must process it to extract its value; this is a major challenge for companies. However, strategies for enhancing the value of information must evolve, because the technologies currently in use do not allow the volume and quality of information held to be fully exploited. “Even the treasures residing in companies’ information systems are undervalued. In this preliminary stage, Artificial Intelligence provides relevant answers and innovations are born”.

Thus, new ways of storing and representing information are appearing and are derived from the Internet but also from research on the human brain. At the same time, new ways of using information are emerging, with information becoming the vector of commerce, as has been proven by players such as Uber in the transport sector. Information becomes dynamic and is made available to the user when and where he needs it. And finally, and perhaps most importantly, systems are able to learn from users to continuously improve.

A market where there are many players, but few are really effective

Today, Artificial Intelligence technologies are mostly public thanks to contributions from private actors such as Google or Facebook, university research centers or open source initiatives. The challenge for companies specializing in Artificial Intelligence is to know how to manage the entire chain: from acquiring internal and external data, formatting it, orchestrating AI processes, publishing the results to users, and finally learning how to use it and the results obtained, and ensuring that it is put into production and maintained in operational conditions. This is precisely oppScience’s strong point, having developed a complete service platform that goes from capture to restitution, where often several technological bricks are needed to meet expectations. “The installation of our platform has replaced three tools within the DCPJ. Our solution allows us to be faster in the processing of information”. The other advantage of the oppScience platform is its automatic language processing engine, which allows us to manage the discovery of relationships and to restore them in a relevant way according to the context of use, following the example of what we know about Facebook or LinkedIn. “In a hyper-connected world, the surveillance of a person and a company can no longer be apprehended without taking into account the ecosystem of their relationships”.

After-sales service, business and competitive intelligence: the multiple applications of AI

If systems only learn what they are shown, they are effective. “Today’s artificial intelligence will not make a car capable of crossing the Place de l’Etoile in Paris; however, it will be able to detect weak signals from a large initial volume of information, whereas humans will consider this information to be insignificant. In this framework, we deal with open data, present on social networks and websites but also internal data. We thus bring a real added value to the after-sales services of companies. In terms of business intelligence, our platform will also be able to detect weak signals regarding customer requests. We also carry out business intelligence and competitive intelligence missions. We cross-reference information published on the Internet with officially published data and information published on the monitored companies. The aim is to anticipate the expectations of customers, whether they are individuals, companies or even states. In summary, oppScience provides a complete solution from the acquisition of information to its availability to users at the right time and in the right context. Our approach allows us to deliver results in agile mode in a few weeks where more fragmented approaches require years”. In this large batch of information, oppScience develops a qualification of each piece of information, treating each one at its true value.
Meeting the expectations of companies requires a training of the algorithm. “We cannot train our algorithms for everything. We have solutions to start the project, then the algorithm is trained by a user; it interacts with the machine through a feedback loop that improves performance over time”. It is essential that a person tells the machine whether its first results are good or not. “Then the improvement is self-sustaining and the optimization is endless. There are two possible feedback contexts. The first is to provide an answer to a question given by a user. The Oppscience platform provides targeted information in a defined context, with a 360° vision. The second is to set up an alert system: in this case, the user is informed of any change in the monitored environment, which is the case during a business intelligence.

OppScience was created in December 2017 by Saïd Rkaibi, CEO of the Medtech group and Gilles André, software entrepreneur. The creation of the company is the result of the consolidation of technologies and know-how from different academic and private initiatives.
Its first offer is bee4sense, a platform for information valorisation; it is a reference solution in the fields of investigative assistance – General Direction of the National Gendarmerie – Economic Intelligence (BNP Paribas) and Customer Support (Schneider Electric).
The experience gained from these first references has enabled oppScience to propose ready-to-use solutions in the fields of investigative assistance and security targeting administrations (internal security, defense, police, justice) in France and in Europe, but also in risk management, economic intelligence and customer support, targeting large companies. Beyond that, oppScience works on various vertical solutions in partnership with business experts and with a partner in the field of conducting and evaluating territorial policies, which demonstrates that information enhancement is a transversal technique with multiple use cases. Moreover, while there is still a budgetary issue that can hinder smaller companies, solutions are becoming more democratic. The company proposes a 4-week approach to demonstrate the reality of the return on investment .
oppScience already gathers a team of 40 people. In 2019 the company achieved a 100% growth of its turnover compared to 2018 and expects the same growth in 2020.