AI in finance is a reality for Criteo

In their room at 9and district of Paris, a team of financiers and scientists from Criteo, a company specializing in targeting online advertising is just around the corner. They have even been working for several years in close cooperation. To better manage the growth of the group, the finance department uses a diary income forecasting algorithm, developed by data scientists. ” We use artificial intelligence as our sales forecast reference data for make revenue forecasting more reliable and eliminate human prejudices. If, for example, a seller underestimates sales in the previous quarter, he will usually be more pessimistic in the coming months. The algorithm allows us to have a objective and scientific point of viewto erase these notions is too ambitious or overly cautiousexplains Pauline Keylard, Head of Financial Planning and Performance Analysis at Criteo.

The tool is essentially objective

The algorithm analyzes reality and provides revenue forecasts and trends more than a quarter based on past events. ” The tool is self-learning. It analyzes tens of billions of data to explain the growth of the country, the customer, and predicts turnover andmost likely sales in the next quarter ยป, reports Jerome Kober, director, product analyst at Criteo. An instrument that is of little interest to financial management. This gives the opportunity to factualize and brings something calm budget discussions between stakeholders. ” The algorithm has an average 2% errors. In general, this is weak.believes Pauline Kayar. This is less when you look very subtly and in detail, because the tool cannot predict details “.

Added value of man

If artificial intelligence can predict a period of advancement in Korea or China, it, on the other hand, cannot don’t expect a global Internet outage or a health crisis. ” In 2020 we have put the tool away because the data was inaccurate. We looked a lot at what was being done in the market guess the curve Covid was going to go through. When we began to realize that the crisis would be much longer, we set about it approach by client and by sector of activity to understand how the turnover was disrupted “- explains Jerome Kober. Hence the interest in take a step back about the information provided and the data to be taken into account. ” AI cannot replace humans. What adds value is this financial work Commented Pauline Kayar. By comparing machine learning statistics with sales and market growth forecasts to see what seems realistic to them, financial teams and the CFO thus contribute the most profitable activity of the company such as process improvement management. “It’s clever piloting offers a diverse and rich playground. We financiers feel privileged to train non-financiers and have access to these conversations at a very high level with engineers. Satisfies Pauline Kayar.


Social reason : Criteo

Responsible for financial planning and performance analysis : Pauline Kayar

Sphere of activity : provider of technological solutions for online advertising

The legal status : HER

Head office : Paris 9

Leader : Megan Clarken

Effective : 2,500 employees worldwide

Turnover : $ 2 billion