FAST LEARNING AI

About FLAI

FLAI – Fast Learning Artificial Intelligence

FLAI is an artificial intelligence able to generate “multilocal” advertising, that means creating advertising messages and displaying them in an optimal & local way for each retailer’s store. FLAI can be controlled by a man-machine interface: the ARMIS SaaS Console.

Scientific challenge

In order to meet the technical and scientific challenge of optimising multilocal campaigns, the development of FLAI had to respond to two structural biases that go beyond the classic framework of machine learning.

  • The first is geographical: the volume of data collected is enormous, but it is concentrated locally around the stores rather than being evenly distributed over the territory.
  • The second is time-related: because physical stores need to communicate over short periods of time, optimization decisions must be made quickly.

FLAI features

01

Circular digitization

FLAI is able to take into account and load the entire print circular and create a database from this input. This “circular feed” is cleaned, structured and enriched by FLAI.

02

Mass creation of ads and campaigns for each store

FLAI can enrich the information in the product feed and then create millions of ads and thousands of campaigns on all the digital platforms (Google, Facebook, Xandr, DV360, Waze, etc.) thanks to deep learning techniques such as visual recognition and Natural Language Processing (NLP) for all devices (mobile, desktop, tablet). These advertising objects (messages, campaigns) are personalised for each catalog and each retailer’s store.

02

Mass creation of ads and campaigns for each store

FLAI can enrich the information in the product feed and then create millions of ads and thousands of campaigns on all the digital platforms (Google, Facebook, Xandr, etc.) thanks to deep learning techniques such as visual recognition and Natural Language Processing (NLP) for all devices (mobile, desktop, tablet). These advertising objects (messages, campaigns) are personalised for each catalog and each retailer’s store.

03

Locally optimized ad delivery for each store

FLAI is able to make decisions using only data from the stores and the current campaigns in order to optimize the ad delivery on multiple dimensions: budget allocation between the different platforms, selection of the best offers, devices allocation, by effectively taking into account the native intelligence of the platforms as well as the multilocal constraint.

A massive data volumetry

personalised ads created for each

store generated per operation

communication opportunities to chose

around each store per campaign