AI assistant for choosing gifts for Media Expert stores.
Choosing a gift is the moment when e-commerce customers most often need support. They know they want to buy one, but they rarely can point to a specific product. Instead, they usually operate with general information — who the gift is for, the budget, and the occasion.
With the very broad product range of the Media Expert online store, this could lead to user fatigue from browsing many categories (“what should I even buy?”), and consequently to postponing the purchase decision.
The goal of the project was to create an AI tool that would take over the role of a shopping advisor and help the user move from a general need to a concrete product recommendation, shortening the path to making a purchase decision.
Project challenge
The key challenge was to design an AI assistant that conducts conversations in natural language while at the same time operating based on clearly defined business rules.
The solution had to:
- understand imprecise, colloquial descriptions of needs,
- translate the conversation into specific selection criteria,
- work exclusively with the current Media Expert product offering,
- maintain speed and stability under high traffic.
An additional challenge was security — both protection against bots and abuse, as well as control over content entered by users, including sensitive data.
Solutions
For Media Expert, we created a dedicated AI gift-selection tool based on dialogue and precise recommendation logic.
The AI assistant uses a language model to conduct conversations and understand user intent. Its task is to translate the user’s language into specific selection criteria by analyzing statements and structuring information such as:
- target recipient group,
- budget,
- occasion,
- gift context.
The collected information is organized into a standardized format that allows the AI tool to work exclusively with the current offering of Media Expert.
Based on this, products are selected from the catalog, and recommendations are ranked according to fit with budget, age, purpose, and conversation context. The tool also ensures variety in suggestions so that users do not keep seeing the same products.
The outcome of the conversation is a set of ready-to-use product tiles with images, prices, and purchase links, along with simple next-step options such as narrowing the selection or changing the direction of the conversation. Responses appear in real time, which shortens perceived waiting time and improves the user experience.
The entire solution was designed so that the most complex decision logic resides on the application side, while AI is used where it provides the greatest value — in understanding language and conducting dialogue.
Results of our work
The implemented AI tool:
For Media Expert, this means a better shopping experience, an AI tool tailored to e-commerce realities, and a solution ready for further expansion with new scenarios, seasons, and campaigns.
Technologies used
Applied technologies, AI solutions, and integrations:
- Azure OpenAI,
- OpenAI,
- JSON,
- AI Guardrails.
See also
about your project
Call or write to us: biuro@sungroup.pl
