Gepard PIM Product Updates August 2025: Smarter Pipelines, AI Shopping Agents & More

August was a busy month here at Gepard – we’ve been working hard to roll out features that make managing and delivering product data simpler, faster, and a lot more powerful. Here’s what we’ve been up to:

AI Shopping Agent Is Here

One of the most exciting milestones this month was launching the very first version of our AI Shopping Assistant. This isn’t just a prototype; it’s a fully working agent that can already chat with real users and help them find products in real time.

We added streaming responses so shoppers don’t have to wait for the entire answer to generate; they see words appear as they’re produced, which feels much more natural and interactive. The agent is also connected to real product data, which means it can recommend actual items from your catalog instead of giving generic answers. On top of that, we focused on making it robust: improved error-handling ensures that if something goes wrong during a conversation, the user still gets a smooth experience.

Behind the scenes, we built a secure back-end-for-front-end (BFF) layer so that sensitive API tokens stay protected, even when you embed the chatbot on external sites. This means you can safely use the assistant on your store, partner sites, or demo environments without worrying about leaking access credentials.

Sergii Shvets

Founder & CEO at Gepard

⚡ Why this matters: Better guided shopping experiences keep customers engaged longer, answer their questions immediately, and ultimately help convert interest into purchases.

Smarter Pipelines With Full Editing

Managing product data pipelines should be intuitive, so we focused this sprint on giving you more power and flexibility. You can now freely add, remove, or rearrange nodes within a pipeline and instantly see the result in a clear, visual flow. It’s no longer necessary to rebuild an entire pipeline just to adjust one step — now you can make changes on the fly and immediately see how data will flow from source to export.

We also added the ability to configure exports directly inside the pipeline, including both one-time exports and scheduled recurring ones. This saves time because you no longer have to switch contexts or configure these steps manually outside the pipeline builder. To prevent errors, we fixed bugs related to duplicate catalogs and connection nodes, so the data you pass through the pipeline stays clean and consistent.

⚡ Why this matters: This level of control means you can react quickly to business needs, experiment with new data flows, and reduce the risk of human error, all while speeding up time-to-market for your products.

Amazon Browse Tree Support

Listing products on Amazon can be tricky when categories are too generic. That’s why we added full support for Amazon’s Browse Tree Structure (BTS) reports. You can now drill down into the exact sub-categories you need, whether that’s “Light Bulbs → LED” or a very specific product type.

Previously, Gepard only allowed the selection of broad Product Types, and Amazon’s algorithms would place your product wherever they decided it fit best. Now you have complete control over where your product appears, leading to more accurate placement in Amazon’s search results.

⚡ Why this matters: The right category means the right shoppers find your product faster, improving discoverability, organic ranking, and ultimately sales.

AI-Powered Product Enrichment

We know how time-consuming it can be to manually fill in missing product features or attributes. This month, we built new endpoints that allow our AI to analyze PDF datasheets and other media files attached to a product, automatically suggesting which features should be filled or corrected.

Instead of your team digging through documents to figure out what’s missing, the system can now highlight gaps and even pre-populate values where possible. This dramatically reduces the manual effort required to maintain a high level of product data quality.

⚡ Why this matters: Rich, complete product data drives better customer experiences and higher conversion rates — and now you can get there faster with less manual work.

Security & Compliance Upgrades

Security and compliance are never “done,” but we made major strides this month. We rolled out a BFF application that safely manages API tokens for chatbots, preventing them from being exposed to front-end code or third-party scripts. We also continued our work toward ISO 27001 alignment: moving from shared development accounts to named user SSH access, rolling out TLS encryption across services, and improving VPN infrastructure.

On top of that, we integrated Gepard with the European Chemicals Agency’s Skip Database. This allows automatic submission of IUCLID I6Z files describing product components and any hazardous substances they may contain. Manufacturers and retailers now have a reliable way to stay compliant with EU chemical reporting requirements directly through Gepard.

Sergii Shvets

Founder & CEO at Gepard

⚡ Why this matters: Your data stays secure, and you avoid regulatory risk — giving you peace of mind and helping build trust with partners and customers.

Faster Testing & Delivery

Slow testing can delay releases, so we put effort into making pipelines more efficient. By re-engineering test execution, we cut average runtime from over an hour to around 20 minutes. We also expanded automated test coverage for imports, exports, and product updates, making it easier to catch problems before they reach production.

⚡ Why this matters: Faster tests mean faster deployments, fewer bugs, and a smoother experience for both your team and your customers.

Smarter Imports & Debugging

For teams working with complex imports, we’ve added the ability to download API response archives. This makes troubleshooting issues much easier, since you can see exactly what data was received and replay it locally if needed.

We also added rate limiting to protect against abuse and improved auto-creation of exports for connectors, making data flows more robust and easier to manage.

⚡ Why this matters: More reliable imports mean fewer surprises, faster issue resolution, and better stability for your entire product data process.

Want to See These Updates in Action?

This was a big month, but we’re just getting started. If you’d like to explore these features or see how they can make your product data operations smoother, we’d love to show you a live demo.

Request a demo or reach out to our team — let’s talk about how Gepard can help you deliver better product experiences, faster.