Internal opinion surveys: opensquare releases collective intelligence
© Fotolia - Denchik
Opensquare, a young start-up originating from the ALMAnaCH project team at the Inria centre in Paris, is reinventing internal opinion surveys. By automating the process of collecting and analysing the responses, the company is gives employees the opportunity to speak and presents their recommendations in a relevant way in order to release collective intelligence.
“Opensquare believes that there is much to take from collective intelligence. Colleagues are a mine of information and ideas that can create or strengthen collaboration and the sharing of good practices. ” This is the general philosophy of the young start-up opensquare, as defined by Benoît Sagot, one of its founders and head of the ALMAnaCH project at the Inria centre in Paris. The structure, which has only existed for six months, specialises in the automation of the analysis of information coming from internal employee surveys.
A researcher in automatic language processing at Inria, Benoît Sagot has developed the IT tool enabling the automation of the collection and analysis of survey results in 55 languages, leaving the business expertise to his three associates, who are specialised in human resources. “Before now, the collection and analysis of the responses was done manually. It was a long and costly process. Add to this the different languages when the company is international, and there were frequent coherence issues. Hence the idea of automating this entire process ", he analyses.
However the tool goes much further. Indeed, as the researcher explains, by cross-referencing the data he can identify very specific issues based on the answers originating from open-ended questions: "If a subject is brought up repeatedly in answer to an open-ended question, opensquare will not only be able to highlight this problem but also measure the importance of this subject for the colleagues. "
Opensquare also aims to carry out the entire survey process and make the questionnaires more relevant thanks to this software: “opensquare is a human resources consultancy firm with innovative technologies and approaches ”, Benoît Sagot notes. Once the questionnaire has been drawn up with the client, it is published online so that the employees can answer it. The start-up then recuperates the data which is processed by the software and automatically generates PowerPoint presentations that meet the expectations of the managers. “These presentations are then the starting point for expertise and consulting work carried out at the highest levels within the organisation ”, the scientist specifies.
A company borne out of Inria's technology transfer
The fact that the opensquare start-up is doing very well (it already has around a dozen clients, half of whom are CAC 40 (French stock exchange) companies) is no doubt due to a very strong collaboration with the Inria centre in Paris. Indeed, it is following in the tradition of other start-ups having benefited from Inria's technology transfer. What Benoît Sagot really likes is having one foot in research and the other in industrial application: “Even if, above all else, I am an Inria researcher, I like to tackle the different demands between research and industry. ” A complementarity that boosts his research work: “They are two very complementary worlds. Industry expectations provide me with invaluable insight on the type of research to carry out. Value creation takes place in both directions ”, he continues.
For the future, managers want to push the automation of the analysis as far as possible, but also go beyond internal surveys. “Today is a time of great reactivity within companies. Internal social networks, shorter but more frequent surveys, annual appraisal meetings or 360 degree surveys will provide us with new opportunities to process other information coming from employees ”, the scientist analyses.
For Benoît Sagot, deep learning (a statistical approach based on neuron networks) is currently experiencing a promising level of success and is progressively becoming the norm in the automatic processing of languages. As a result, we could go further than just language and add to it all of the non-linguistic context: "A necessary tool to better exploit the mass of data circulating on the Internet, such as text, images, video, etc. ", the research specifies.