Success story II

  • Home
  • /
  • Success story II

The challenge

Develop a software tool to provide Commercial Intelligence to a silicon producer selling all over the world.

This project describes a case in which Artificial Intelligence is used as a tool to analyse and predict the behaviour of a determined market.

Why and what for?

The globalisation of markets has led to substantial changes in companies’ business models. Today the market is the world and companies are drawing up their growth strategies considering potential suppliers or customers in the five continents. The opportunities this model provides are nevertheless accompanied by a number of threats.In any market scenario, knowledge of its parameters is critical for making the right decisions. Our solution, used as an engine to analyse and predict the market’s behaviour, provides intelligence and documented arguments for making the right decision.

In this success story, we check how using Artificial Intelligence helps a multinational producing and selling silicon to keep an eye on the behaviour of the global market in which it operates.

How?

Commercial intelligence. This is the value we provided to this multinational for the consideration and making of its business decisions. First of all, we identified public data that different international bodies publish at a determined regularity. As always, the data is there, what we need to do is gather it together and analyse it as a whole to extract from them high added-value knowledge hidden beneath a chaos of figures and variables.

Once all the sources were identified, the big data technology we incorporated managed to: order, categorise, clean and integrate all this large quantity of information. Once this phase was completed, we designed a scorecard to show its results clearly and intuitively, answering a set of queries previously defined with the customer. Market flows, imports and exports from the different countries, quantity of tonnes of the operations, sums at stake, time periods, geographical areas…. This and other relevant information (exchange rates or identification of deviations) was not only integrated historically, taking data from the last few decades, but is also updated automatically on the customer’s dashboard with all the data being made public.

After monitoring everything that happened in the past (interpreting trends, outstanding operations, milestones, etc.) and automating the updating of all the sources of value, the moment came to apply automatic learning technology to provide information on future scenarios. To do this, as in all our projects’ prediction phases, we elaborated a great number of algorithm combinations, until we identified the most efficient one to respond to the situations considered. We trained the model with historical data and checked its validity to start to apply it with real-time data.

Conclusions.

Currently, our customer immediately knows all the data that affect its market, can answer a large number of queries relevant for decision-making, has access to predictions of future behaviours with an appraised margin of error. In short, it has a better knowledge of its position in the market and has easy and intuitive access to a large quantity of information without the need to try to extract knowledge from the personal, periodic analysis of a chaotic data scenario.