In order to reinstate the salesperson of today, soyooz has developed a tool based on artificial intelligence (propriety technology) to reproduce the kind of intelligence and know-how used by a salesperson or that of a product specialist.
The initial stage: Modelling human expertise
In exactly the same way a person is trained to sell a particular commodity, the virtual salesperson learns everything there is to know about any one or a range of products.
In order to do this, each product is decomposed, analysed and simulated so that the technical characteristics can be associated to consumer usage.
The commercial strategy of our clients is also taken into account in this simulation.
Standard recommendation methods are generally based on a form of filtering: strict criteria (e.g. price range) or graduated (e.g. from beginner to expert…).
However, the act of purchasing is rarely that simple and more often than not the fruit of compromise. . So to characterise a product in a binary manner (filtered or decision tree) can only have limited results that are often both frustrating and misleading. (For example: the consumer could miss out on the ideal product simply because the price is but 2€ over and above his/her budget).
Our calculation engine works completely differently by evaluating the already mapped out products and user criteria.
It does not eliminate any of the products and functions, like the human brain it establishes the best compromise possible, and this, in less than one second… even for catalogues containing thousands of products.
The observation and analysis of consumer interest given to the recommendations allows us to constantly improve modelling the choice of criteria and recommendation process (machine learning).
Soyooz has developed a unique ability in terms of recommendation orientated user experience. Soyooz has a thorough understanding of how to communicate in order to efficiently and rapidly detect consumer needs.
More than 90% of customers who use our solution are pleased to discover products that are well adapted to their needs.
This performance requires a thorough understanding:
- Consumer habits (how to attract the customer’s attention? how to formulate the questions? etc.)
- Building up the chain of questions to be asked (How many questions are needed? How to react to a response? Which questions not to ask? etc.)
- Presenting results (What information should be delivered? How many products should be presented? etc.)