Client: Astra Polska
Case study
Astra Polska – digitization of the product catalog using PIM and AI
Range
Akeneo PIM implementation, eCommerce consulting, data model creation, PIM and ERP integration, data modeling, data migration, project management, AI-based automation
Astra Polska is a renowned Polish manufacturer of leather goods and art supplies, whose products have been found in homes, schools, and offices across the country for years. The main reason and motivator for the change was the scattered data across multiple locations—from retail price lists, through marketplace platforms, to the website—which hindered consistent and effective management of the offering.
In response to these needs, we emerged as a strategic partner, aiming to digitize and centralize the product catalog. As part of the project, we implemented a PIM system, which became the central source of product truth for the entire organization. Using AI tools, we automated and streamlined processes that previously required significant amounts of manual work.
Centralization of product data – challenges and goals of PIM implementation
Our primary strategic goal was to create a single, coherent source of product data that would serve as the foundation for e-commerce development and collaboration with key business partners. To achieve this, we defined the following operational goals:
Creating a scalable data model
Developing a solid model that will support the work of not only the eCommerce department but the entire organization, enabling easy management of product attributes, categories and variants.
Centralization and digitization of the product database
Collecting scattered product data in one place and giving it a structured, digital form.
Advanced integration with the ERP system
Building two-way integration between the PIM system and Comarch XL, taking into account specific customer needs, such as dynamic management of the content of product sets (e.g. school supplies) and automatic updating of their composition.
Improving cooperation with B2B partners
Using PIM as a central data source that can feed the systems of the largest clients, such as large retail chains, ensuring they always have up-to-date and complete information.
Automation of work with content
Implementation of artificial intelligence-based modules to automate translation processes and generate product content.
Implementation and training of the team
Preparing the Astra Polska team to work independently and effectively with the new tool.
What does the PIM system implementation process look like?
Project implementation stages
The implementation process was divided into four key stages that ensured smooth and effective achievement of the intended goals:
01
Product data analysis
We began the project with an in-depth analysis of the company's existing data resources, identifying their sources, structure, and key attributes.
02
Preparation and implementation of the data model
Together with the Astra Polska team, acting in an advisory role, we developed the target data model. We reviewed the category tree structure, attribute naming, and their values to ensure the model was as intuitive and scalable as possible.
03
Integration of ERP Comarch XL with PIM
We built a dedicated integration with the Comarch XL system, which automated the flow of product information, including handling complex sets with variable content.
04
Data processing and migration using AI
Using advanced AI models (including Claude 3.5 Sonnet), we generated batch files for PIM. Based on existing descriptions in the XLSX files, the AI automatically mapped and assigned appropriate attributes to the created data model.
Have any questions?
Let's talk about cooperation.
Using AI in the PIM system – automating work and improving data quality
What sets this project apart is the deep integration of artificial intelligence into the data management process. Our efforts weren't limited to simple automation—we treated AI as a partner in the analytical and implementation process. What did our client gain from this?
Automating work with data and reducing human work
Using AI to analyze and allocate attributes from XLSX files to the target data model significantly shortened the time and reduced the costs of migration, eliminating the need for many hours of manual work.
Intelligent content modules
We have implemented AI modules that not only automate the translation of attributes into various languages, but also support the generation of unique product descriptions, which is crucial for SEO and UX.
Ensuring data consistency and quality
AI proved invaluable in detecting illogicalities in existing data. For example, the algorithms were able to identify situations where a product's title and attributes indicated green, while the description suggested orange.
Enriching the data model
Artificial intelligence also served as an advisory tool, indicating potential additional attributes and values that should be added to the model to better describe products and respond to customer needs.
The effects of implementing PIM with AI – the foundation for eCommerce development
The implementation of a PIM system using artificial intelligence at Astra Polska is an excellent example of synergy between advanced technology and strategic consulting. The project's success was possible thanks to close collaboration with the client and our involvement as an advisory body at every stage – from data model conception to implementation.
This project demonstrates that digital transformation in product management is not just about implementing a new system. It's primarily about a well-thought-out strategy, a well-designed data model, and the bold use of innovative tools like artificial intelligence. For companies planning similar implementations, we recommend choosing a partner who not only provides technology but also actively advises and supports process optimization, ensuring that the investment delivers real and measurable benefits.
Have any questions?
Let's talk about cooperation.
Do you have an idea, a ready specification or a business need?
Write to us. We will contact you within 24 hours.