AI-powered product search for a beauty e-commerce platform

Service
: AI agent

Industry:
 E-commerce, retail

Project type: 
AI-powered product search, customer support

Core focus
: Semantic product search, personalized recommendations

Project duration:
 4 months

About the client

GALERIJA BALTIK is one of the leading distributors of professional beauty products in Latvia and the owner of the e-commerce platform https://salonline.lv. The platform serves both individual customers and beauty professionals, offering a wide range of professional cosmetics, accessories, electrical devices, and equipment.

Challenge

As the Salonline product catalog grew to over 35,000 items, it became increasingly important to ensure faster and more intuitive product discovery.

The existing search functionality effectively supported searches by product name, keywords, and codes. However, with the growing assortment and increasingly diverse customer needs, there was a clear opportunity to significantly improve the search experience using AI.

Private customers increasingly search based on needs or desired outcomes, such as “shampoo for dry hair” or “hair color for gray coverage”, rather than specific product names. At the same time, professional users require fast and precise product search by code.

As a result, there was a need for a new search experience that could better understand user intent and help customers reach the most suitable product more efficiently.

Business needs:

·       Faster product discovery

·       More accurate search results

·       More intuitive product selection

·       Reduced workload for customer support

·       A more efficient path to purchase

Project goal

The goal of the project was to enhance the Salonline e-commerce platform with an AI solution that helps customers navigate a large product catalog more efficiently and make more confident purchasing decisions.

Solution

ZenIS developed and integrated an AI-powered product search agent, “SALON LINE PRO Assistant,” into the salonline.lv platform. The solution enhances traditional search with a conversational, intent-based product discovery experience.

Unlike traditional keyword-based search, the AI agent analyzes the context of the user’s request and understands its meaning. Users can describe their needs in natural language, while the assistant asks clarifying questions when necessary and provides a focused selection of the most relevant products - significantly simplifying the decision-making process, even when the exact product name is unknown.

What the AI agent provides:

AI-powered product search

·       Understands user intent and search context 

·       Supports natural language queries 

·       Asks clarifying questions to improve accuracy

Personalized product recommendations

·       Selects the most relevant products for each situation 

·       Provides direct product links 

·       Shortens the path from need to purchase

Support for different customer segments

·       Tailored experience for both B2C and B2B users

Integration with product data

·       Real-time access to up-to-date product information 

·       Integration with the existing product catalog 

·       Automatic data synchronization

Brand-aligned AI assistant

·       Responses aligned with Salonline’s communication style 

·       Custom-built conversation scenarios 

·       Interface and design adapted to the brand

Implementation

The implementation started with a simple but critical question: At what point do users get lost in the product search process?

The analysis showed that the issue was not a lack of data, but rather the opposite, too many choices without clear direction. Instead of adding another layer of filtering, the focus shifted to a fundamentally different approach: conversational search.

The AI agent was integrated directly into the existing e-commerce environment, without disrupting the familiar user experience.

The implementation included:

·       Analysis of product search and customer journey

·       Integration of the AI agent with the website and product database

·       Data synchronization

·       Configuration of recommendation logic

·       Development of conversational scenarios

·       Product search by code functionality

·       UX/UI adjustments

·       Testing, usage analytics, and performance monitoring

·       Ongoing technical support and consulting

Project results

The impact was immediate, not in terms of features, but in user experience.

Product discovery shifted from browsing for several minutes to a short, guided interaction. Users can move from a general request to relevant product recommendations in approximately 30 seconds.

The most significant improvement was in simplifying choice. Instead of navigating hundreds of results, users receive a focused selection of the most relevant products, reducing cognitive load and accelerating decision-making.

The solution also transformed the search logic itself — from keyword matching to meaning-based understanding. This enables the system to process real, human-like queries and deliver more accurate recommendations. Importantly, the solution works equally well for different user segments - both end customers searching by need and professionals requiring fast, precise product lookup by code.

From a business perspective, this results in reduced customer support workload and a more efficient path to purchase.

The AI solution was designed as a modular and scalable platform, allowing it to evolve alongside business needs. Future development may include additional customer and behavioral data integration, advanced B2B scenarios, proactive communication, and expansion across multiple customer interaction channels.

Key benefits:

Faster product discovery

From minutes to seconds, users quickly reach relevant results 

More accurate recommendations

Fewer irrelevant results, more precise matches 

Simplified decision-making

From hundreds of options to a curated selection

This project demonstrates a fundamental shift in e-commerce, from product search to product consultation. When a system understands customer intent rather than just keywords, the selection process becomes faster, simpler, and more confident. That’s where real business value is created.

Technologies used

AI Agent | Natural Language Processing (NLP) | Product Database Integration | Semantic Search | Keyword Search | Product Recommendation Logic | Cloud Infrastructure | Conversational User Experience | E-commerce Integration | Usage Analytics | Data Synchronization

We found our collaboration with ZenIS professional and results-driven. The team quickly understood our business needs and translated them into a practical AI solution. We particularly valued their flexible communication, technical expertise, and ability to seamlessly integrate the solution into our existing e-commerce environment, significantly improving the product search and selection experience for our customers.”

Armands Bulkovskis
CEO, GALERIJA BALTIK