AiRS allows to optimize processes and increase performance based on a seven step learning process. This process is based on the correlation of the latest visitor data collected with all previously collected visitor data.

Step 1 - Content Object Classification
The initial step when using AiRS is to determine a classification grid, including objective and subjective content attributes, and to classify products, banners, text links or other content objects to be recommended by AiRS.

Step 2 - Visitor Exposure
The AiRS closed-loop self-learning process begins when a visitor connects for the first time to a web site and gets exposed to a set of content objects. Content objects can consist of any type of offer, including editorial information, sponsored text links, commercial offers, promotions, advertising banners or search results..

Step 3 - Data Acquisition, Filtering & Analysis
In step three, AiRS records the actions and inactions of a visitor based on the content he has been exposed to. AiRS then analyses how these behaviours and interests compare to those of all other visitors.

Step 4 - Visitor Interest Maps & Audience Segmentation
In step four, AiRS projects all visitors onto several two dimensional interest maps based on their DNA attributes. These projections allow to determine micro audience segments of visitors who share common behaviours and interests and are likely to have similar responses when exposed to similar contents.

Step 5 - Content Object Recommendations
AiRS can now match content and visitor DNA's to determine which contents to show on the pages a visitor calls up during a session. AiRS operates based on a dual online and offline process to achieve the best possible results in the shortest possible time. During sessions, AiRS integrates the latest known visitor interests and non-interests to determine on the fly what to recommend to the visitor at that given moment.

Step 6 - Delivery
AiRS delivers the selected set of contents at a given time, in a given context, to a given visitor. This selection is determined by the recommendation engine and fine tuned with the use of filters activated by the operator. Once content objects have been delivered to the visitor, a new "learning" cycle begins. Each time this cycle is repeated, assumptions and results improve. AiRS has a built-in set of alarms to immediately inform operators about products that are selling above or below expectations and to monitor inventory.

Step 7 - Reporting & Analysis
AiRS provides marketers with an in-depth understanding of their audience. The AiRS reporting tools allow to identify "profitable" and "non-profitable" visitor segments and emerging trends. With this foresight, one can adjust offer to demand and introduce new product assortments or, for example, adapt the pricing strategy. AiRS "learnings", encapsulated in each visitor's DNA, allows to establish direct correlations among hundreds of visitor attributes. AiRS provides a wealth of audience information to marketers unmatched by traditional marketing research methods.