Since its creation in 2015, the Swiss Predictive Layer has come a long way. It has clients such as Axpo (energy) or ArcelorMittal (industry) and its technologies are even used by certain States to work on their energy mix and its development.
Predictive Layer’s business sector? Generate forecasts on time series. “Our job is to predict the future of our customers regarding everything related to supply and demand”, summarizes Serge Rigori, CEO of Predictive Layer.
AI and data
The company has developed an AI-enabled offering, which is able to draw on historical company data, but also third-party datasets, which Predictive Layer acquires from key market players (e.g. AXSMarine or Kayrros).
The company has also worked hard on another aspect, the explainability of its forecasts. “Companies need to understand the rationale of the forecasts we offer them, before giving them credit. For our forecasts to become a real decision-making tool, the explanatory dimension of the forecast is as important as the forecast itself. »
Predictive Layer relies on around thirty specialists with an important background in mathematics and physics, with regular collaborations with research laboratories.
Short or long term forecast
Predictive Layer works on two levels: a long horizon, with forecasts over several years that allow strategic production planning; and a shorter horizon, in order to optimize the supply chain according to demand.
“Our users are either companies that want to use our tool completely independently, or organizations that simply want to consume turnkey delivered forecasts,” continues Serge Rigori. In some cases, for example the reduction of CO² emissions, we can even propose an order book to be carried out to optimize emissions. »
The company therefore sells both a solution and an expertise. In the first case, Predictive Layer delivers its software and trains users, so that they become autonomous on the solution, while in the second, it provides forecasts, possibly supported by action plans. “In both cases, we are constantly benchmarked by our clients, who will compare our forecasts with reality. »
Note that for its software, Predictive Layer does not impose SaaS mode. “The information handled by our solution is often essential and critical for our customers, who prefer the solution to be installed on site and not in the cloud. »
An original approach to pricing
Predictive Layer also helps companies set their prices. “We are not the first players to tackle the subject of pricing, but our expertise allows us to adopt a different approach, explains Serge Rigori. We do not rely on quantiles, because we consider that the price is only the consequence of supply and demand. It is therefore on these two aspects that we focus when setting prices. »