Recommendation Systems, AI-Based Product Search
The major market statistics-based recommendation systems have either become very expensive or have been phased out (e.g. YUSP). Therefore, in response to our customers' needs, we have developed our own recommendation system.
I. In the first step, we created a solution based on customer data analysis, operating on our own algorithm. Our recommendation system takes into account various factors for crafting the appropriate offer, such as previous purchases, purchases of customers with similar behavior, analysis of the current market environment, which examines which products are selling during the given period. Several reports are available about the service, allowing tracking of how the recommendation module contributed to the orders/revenues of the online store.
The pricing of the service consists of two elements:
- One-time installation fee (design, display of recommendation interfaces, data collection)
- Monthly fee: depending on traffic (number of offers, etc.)
The service is available from the beginning of January 2024.
II. In a second step, it is planned to add the AI-based operation or the AI-based product search to the statistical model.
AI-based correction of misspelled search terms
With our OpenAI-based misspelled search term correction module, specific to online stores (e.g., games, dog food, clothing), misspelled search expressions can be corrected. Tests show that it can correct 80-90% of misspelled search terms. To measure effectiveness, we provide a range of statistics to assess how AI-based user misspelling correction has contributed to orders, revenue and user experience.
To enable our customers to optimize their costs associated with AI (tokens) efficiently, we offer the option for clients to configure whether all search expressions entered by users are evaluated based on AI algorithms or only those for which traditional search solutions were unable to return results. We have developed efficient optimisation solutions to reduce the costs associated with the use of AI.
The service is available to our partners for whom we have developed a webshop. Costs are primarily determined by the average monthly number of search expressions after the one-time implementation. The service is available from the beginning of January 2024.
PWA
Progressive Web Apps (PWA) are fundamentally web-based applications that behave like native applications.
PWAs combine the best features of traditional web applications and native mobile applications. They are built using web-based technologies such as Javascript, HTML, CSS, etc. and are accessible through a web browser. The primary advantage of such applications is that there is no need to develop a separate native application for mobile devices, including the entire business logic and discount systems. It is sufficient to make the front-end of the web application PWA compliant , which can be implemented cost-effectively and allows for much faster operation while retaining the benefits of native mobile applications.
Nowadays, more and more consumers use mobile devices, expecting an instant user experience. Therefore, it is not surprising that online retailers are turning towards Progressive Web Apps to meet the growing expectations of their customers. It is very important to emphasize that this technology is not suitable for every client (similarly to native applications) and is initially recommended as an addition to the existing web interface.
We exclusively offer this service to our existing clients or in conjunction with the development of a completely new web store created by us.
Cooperative partners
