Predictive analysis serving industry

Using artificial intelligence and machine-learning, Vaud-based company Predictive Layer processes its clients’ data with that available via open source to make near-term forecasts.

Using artificial intelligence and machine-learning, Predictive Layer processes the data of its clients in the energy, industrial, distribution, and transport sectors, correlating them with open-source data then creating accurate and reliable near-term forecasts.

The system also works in other sectors, with the company providing accurate forecasts for a European retailing giant on frequentation and sales per product for the coming hours, days, and weeks, taking into account various parameters such as weather and vacation periods as well as local events and other marketing promotion calendars. "The stores can thus optimize allocations of staff, shelves, stocks of perishables, and supply logistics," says CEO Olivier Cognet.

Forecasts in real time

For airports and other public transport systems, Predictive Layer carries out automatic forecasts in real time on traffic at the different stages of access, from parking through check-in, security, and shopping areas to the boarding gates. Again, the company automatically takes into account external conditions such as the weather, vacations, and events in neighboring cities, as well as the latest changes in timings and delays in other airports. In short, via a monthly subscription service, Predictive Layer sells its customers a forecasting service that is automated and contextualized, and which includes various types of external local data.

The start-up’s main innovation is its machine-learning algorithm for time series, as well as a simple computerized system with automatic access to available open-source data.

European markets

Of the competition, Cognet mentions the big IT players (IBM, SAP, and Google) which offer generic platforms, but adds that "the core of the machine-learning does not have the same levels of automation or the means of simplified industrial production. Companies that are more specialized in the industrial sectors model and forecast the evolution of their systems, but without fully taking external data into consideration and without machine-learning functions. We use machine-learning for time series, automatically complementing and improving their forecasts by 30 to 50%."
Predictive Layer was founded in March 2015, and includes amongst its clients the Arcelor Mittal group, the industrial services of the City of Grenoble, divisions of the EDF group, and other customers in Germany, Benelux, the United Kingdom, Spain, and Singapore. In the years to come, the company – which has five permanent employees and a network of researchers and strategic advisers in Paris, Zurich, and San Francisco – is looking to concentrate its efforts on markets in Europe by continuing to target the energy, heavy industry, transport, and distribution sectors in particular.

A valuable support from the Economic Promotion

Predictive Layer, originally a French company, specifically elected to set up business in the canton of Vaud. "The impressive density of innovative companies played a decisive role in our choice," says CEO Olivier Cognet.

Since it was created in 2015, Economic Development – Canton of Vaud (DEV) has provided Predictive Layer with various contacts, in particular with trusts, lawyers, the banking sector, the HEIG-VD (School of Management and Engineering Vaud), and the Y-Parc (technology park in Yverdon). A project in conjunction with the Commission for Technology and Innovation (CTI) was set up after Innovaud and the Alliance national consortium for innovation put the company in touch with the HEIG-VD.

The Office for Economic Affairs and Innovation (SPEI) provided various types of financial support to the Vaud-based start-up, in particular by way of participation in international events. The company took part in the 2016 Swisscom Event held at the SwissTech Convention Center of the EPFL. In October of that year, it also made a presentation during the annual event held by Economic Development – Canton of Vaud.

"DEV, Innovaud, and SPEI gave us advice on how to structure a company in the canton, as well as good contacts among those involved on the business and the support side," says a delighted Mr. Cognet. "Exchanges with education and the research sectors enabled us to continue to develop the heart of our innovative CTI project. We also benefited from access to important contacts in the region to develop industrial synergies."