Complete product cards without having to search for data manually

Energokomplekss offers thousands of electrical engineering products from numerous manufacturers, and new items are added regularly. Filling out each product card with a description, technical specifications, images, and documents required a lot of manual work. The company needed a solution that would collect this information, organize it, and prepare it for import.

Energokomplekss.lv Produktu datu automatizācija ar mākslīgo intelektuEnergokomplekss.lv Produktu datu automatizācija ar mākslīgo intelektu

Tested and refined in collaboration with the client

We developed the solution step by step and tested it at each stage together with the client. First, we tested it with individual product groups. The client evaluated the results by category, and after making improvements, we expanded the solution to the entire product range.

  • Analysis of Data Sources and Product Group Attributes
  • Creation of a data structure by product category
  • Prototype development and testing
  • Data Collection from Multiple Sources
  • Identifying conflicting data across sources
  • Generating descriptions and selecting images
  • Customer Verification Process
  • A routine process for handling new products
  • Integration with the ERP system and the B2B online store

Diverse Data, Unified Approach

Each product group has its own technical specifications, and manufacturer data comes in different formats and volumes. The solution had to identify the precise information and populate only those fields that are relevant to the specific product. At the same time, the company needed to maintain control over what data appears in the online store

Challenge: Incomplete and Inconsistent Data

Not all manufacturers provide images, data sheets, or certificates, and the specifications for the same product can vary across different sources. Product descriptions could not be general; they had to be based on facts. Since data for a single product can be contradictory, we had to decide which source to trust and when human verification was necessary.

Solution: Structured data collection using AI

Based on the product code and product group, MI determines which parameters need to be filled in and searches for them. Data is collected in a specific order: first, the manufacturer’s data, and only if that is insufficient, other online sources. MI prepares a description in Latvian and English in accordance with SEO requirements, selects the most appropriate product image, and adds links to data sheets and certificates, if available. Any empty or incomplete fields in the prepared file are identified and filled in.

Automatic Validation and Human Review

Each value entered into the MI is accompanied by a confidence rating and a brief explanation of the source of the information. Entries with high confidence are automatically forwarded. If the confidence level is low or the sources do not match, the entry is sent to the client’s team for manual review. Each entry passes through the statuses “Under Review,” “Ready for ERP,” and “Imported,” so you can always see which stage it is in.

Before importing, the solution also checks whether the product group matches the one defined in the ERP system to ensure the data is placed in the correct category.

Integration and New Products

The verified data is imported into the ERP system, and from there it is synchronized with the B2B online store. On the product page, the buyer is informed that the information has been collected and optimized using artificial intelligence. The solution is designed as a regular process, so new products are processed according to the same principles as the existing product lineup.

Produktu datu automatizācija ar mākslīgo intelektu

Marina Tribocka

SIA “Energokomplekss” Biznesa atbalsta un attīstības vadītāja

“Excellent project management, professional support, and a great attitude. The project was complex, involving both ERP integration and an AI solution for product data, but throughout the entire process, we always knew what was happening and what the next step would be.”

Other projects

An AI voice agent for customer appointments

Automated Call Handling and Customer Appointments

A voice agent, who answers calls at any time of day or night, schedules appointments for clients and transfers them to a representative when necessary.

  • Bookings, rescheduling, and cancellations during a call
  • Reminders via email and WhatsApp
  • Call recordings, call transcripts, and daily reports
  • Dashboard with full control over MI agents
Learn More

SKY & MORE

A functional, user-friendly ecommerce site for premium brands

As e-commerce trends evolved, a premium clothing retailer that previously relied on an informational website needed to expand into online sales. The goal was to create a modern e-commerce platform that supports direct purchases while providing customers with a smooth and convenient shopping experience.

  • Concept development and design
  • Development
  • Warehouse system integration
  • Creation of complex product filters
  • Maintenance and development
Learn More

Sazinieties ar mums

Mūsu komanda priecāsies atbildēt uz visiem Jums interesējošiem jautājumiem. Aizpildiet kontaktu formu, un mēs ar Jums sazināsimies.

Default contact form (EN)

eComStrive - Sazinieties ar mums un piesakiet bezmaksas konsultāciju
<h3>Thank you, we have received your message!</h3>
<p>Our specialist will contact you shortly to arrange a convenient time for a conversation.</p>