Product data is the fuel of modern Ecommerce—but in most companies it’s still refined the old way. Here at
Pimics, our aim is to change that with our
PIM solution built on Microsoft Dynamics 365 Business Central. In this article, we take a closer look at our exciting collaboration with Mendel University in Brno. With the support of European Union funding, we are jointly developing
two AI agents that will operate on Microsoft Dynamics 365 Business Central:
the Pimics AI Agent, which processes product data to streamline product information management, and the
DLD AI Agent, which optimizes warehouse logistics and distribution.
AI-agents technology is advancing rapidly and reshaping industries worldwide. This article explores what’s driving the shift and why we decided to build our own AI-agent solution.
AI agents are on the rise! And why is that?
Let's take a quick look at what was happening in March 2025. At that time Microsoft introduced its first Business Central agents:
Sales Order Agent and
Payables Agent, and the conversation completely shifted from “AI helps me” to “AI works with me.”
We’re moving beyond “ask and wait” assistants (where every output depends on the perfect prompt) toward a scalable digital workforce that can operate with far more autonomy: interpreting data, evaluating context, and taking action! If you want a quick refresher on this shift, we covered it more clearly in our earlier article:
From Copilot to Autonomous Agents: AI in Product Information Management.
Let the numbers speak for us!
IDC—a leading global company that has been analyzing technology trends and opportunities for more than 50 years, shared its predictions
for the development of AI agents. According to IDC,
the number of active agents is expected to grow from approximately 28.6 million in 2025 to:
- 1.3 billion active AI agents in 2028
- 2.216 billion active AI agents in 2030
Fortune Business Insights also highlights just how fast the AI agents economy is accelerating. In its
AI Agents Market Size, Share & Industry Analysis report (published March 2026), the firm reports the following key market-growth figures:
- Global market size 2025: USD 8.03 billion
- Global market size 2026: USD 11.78 billion
- Global market size 2034: USD 251.38 billion
- Compound Annual Growth Rate (CAGR 2025–2034): 46.61%
Grand View Research, Inc. adds another strong signal: the global AI agent market is projected to hit
USD 50.31 billion by 2030, growing at a
45.8% CAGR (from 2025 to 2030).
The growth and adoption of AI agents is undeniable, and their impact is already being felt across industries. To thrive in this evolving landscape, companies and their leaders must now actively integrate AI into their operations.
Hybrid Teams are coming!
How AI Agents reshape workflows and ownership?
AI agents are already reshaping our everyday work, from drafting and analysis to operational tasks. And as more organizations adopt them, workplace dynamics will change quickly. In the near future, we can expect that most businesses will run hybrid teams where people and AI agents work side by side, each doing what they do best.
To clarify the impact of AI on workflows and ownership while managing product data, we will present three illustrative examples:
one from the past, one from the present, and one from the near future!
Past: Managing product data without a PIM
Imagine a retailer onboarding 2,000+ new SKUs from multiple suppliers. Product data shows up everywhere: Excel files, PDFs, emails or shared drives. Someone imports what they can into the ERP, fills the gaps in spreadsheets, renames images by hand, and copy-pastes product descriptions.
Without a PIM tool and well-defined data strategy, every correction slows the processes down, increases the chance of mistakes, and creates yet another version to track. At the end,
publishing only relevant, high-quality product data becomes a struggle, often
leading to inconsistent listings, lower customer trust, and lost sales.
Present: Managing product data with Pimics!
With Pimics, all product information is managed in one place, from importing and enriching to validating and publishing. Centralizing product information helps you stay competitive in fast-moving markets,
reduce the risk of operational errors caused by manual processes, and
accelerate time-to-market when launching or updating products.
- Import: Standardize supplier data into a consistent structure (attributes, variants, categories) to eliminate format friction. With our smart AI-based PDF import, our clients can simply upload supplier documents into Pimics and let the AI extract the right data automatically.
- Enrich: Maintain a single source of truth for all your product data—technical specs, commercial data, translations, and media. With features such as AI-generated marketing texts and product descriptions, AI suggestions for structured attributes, and built-in AI translation, our clients can prepare product data dramatically faster.
- Validate: Use completeness rules and quality checks, so teams see what’s missing before anything goes live.
- Publish: Distribute consistent, channel-ready product data to e-commerce, marketplaces, and partners—without retyping.
If you would like to find out more about the AI functionalities you can already use in Pimics, take a look at our previous article, where you can explore each one in more detail:
Pimics in 2025: Year of New Features and AI Innovation!
Right around the corner: Our own AI Agents!
This is where things become truly future-ready. In collaboration with
Mendel University in Brno,
we’re developing our own two AI Agents: Pimics AI Agent and DLD AI Agent.
The
Pimics AI agent will focus on
transforming product information management, and the
DLD AI agent is designed for complete
optimization of warehouse logistics and distribution.
These two solutions can be implemented together for an integrated approach or separately depending on your business needs—making them highly flexible for any organization working with thousands of product records for Ecommerce, while also streamlining distribution and logistics.
In this article, we will
shine the spotlight on the Pimics AI agent solution and explore its capabilities in greater detail.
Pimics AI Agent: All the details you need to know
Pimics AI Agent will be a tool for automating the intelligent transformation of product data used in product presentations on websites, in catalogs, or in sales offers. This technology will enable the analysis of text inputs, the generation of marketing descriptions, the interpretation of emails, and context-based image processing.
Pimics AI Agent will remove the barriers to digital transformation
With the Pimics AI Agent, teams stop wrestling with messy supplier inputs and start running a true
hybrid workforce—where an
autonomous agent handles the repetitive and error-prone work while people stay focused on decisions, exceptions, and growth.
Our goal in developing the Pimics AI Agent is to address the major barriers to digital transformation, such as:
- the heavy reliance on manual data processing,
- the challenges of working with unstructured information,
- and the limited availability of tools for non-technical users.
By leveraging cutting-edge technologies in natural language processing (NLP), generative modeling, and classification, the Pimics AI Agent will be able to understand straightforward commands given in plain language and seamlessly operate within the Pimics environment, handling the complexity of today’s product data.
The main benefits of the Pimics AI agent include:
- Transforming natural language text into structured data based on natural language instructions
The Pimics AI Agent helps our clients to automatically process text information from websites or brochures. Using natural language processing (NLP), it analyses available texts and extracts specific product attributes, storing them in the Pimics database for future use.
- Generating and modifying text from structured data based on natural language instructions
Thanks to the chat interface, users will be able to generate marketing texts for products in bulk. Users simply select the products and provide instructions.
- Enriching data with context based on natural language instructions
This feature will allow users to use natural language to instruct the agent to check the relationships between attributes based on the specified criteria. For example, it can detect a common issue where an attribute is not listed in the required format.
- Advanced image handling using natural language instructions
This feature allows users to instruct an agent to add to and modify content typically found in digital asset management systems using natural language.
- Natural language–driven selection of multiple records
This feature streamlines many frequently performed user operations. When making price changes, defining a product segment for publication or performing similar tasks, users often select groups of products by exporting them to Excel, where additional calculations are performed and criteria are defined. This part of the process will be replaced by a dynamic view of the data, which can be modified using natural language.
+ Bonus tips for you:
If the topic of AI in product information management is on your radar, you might also enjoy some of our earlier blog articles. You can find the links below:
Are you prepared for the next AI-generation of PIM systems?
We would be delighted to partner with you!
Let's connect!