Whisper founders Martin Kučera and Vojtěch Paukner have been friends since kindergarten. One is an elite programmer; the other is a product manager with experience at companies such as Rossum. They share a favorite question: How could real-world tasks be done more simply, so people don’t have to struggle so much to get things done? In other words: how can activity XYZ be automated?
With the rise of LLMs, the step toward launching an automation-focused startup was a small one. They also had an ideal testing ground for proving they could build a truly useful product: their joint project, the popular charity platform Milý Ježíšku. Coordinating thousands of donors and dozens of children’s homes to ensure gifts arrive on time requires enormous amounts of communication and clicking through a wide range of systems.
“A large share of donor communication involves messages that require complex problem-solving, not just a simple reply. At the same time, it’s important for us to preserve a human touch. Whisper’s pilot operation saved us two-thirds of the time compared to our previous setup. We fine-tuned the agent on the fly, and the data shows that with the current setup we could reach up to 90% time savings. Crucially, we didn’t have to write a single integration for our custom admin system. Whisper only received API access and documentation — it learned the rest on its own."
Whisper was particularly helpful in cases where donors needed to modify multiple reservations at once or when it was necessary to verify information with a children’s home. After validating Whisper in practice, the founders are now bringing the same principle into e-commerce.
In customer support, writing the email itself isn’t what takes the most time — it’s everything around it. To resolve a single request, a person often needs to open five different windows: find the order, verify shipment status with the carrier, notify the warehouse, or update a shared spreadsheet. This “running around” between systems and actions within them is exactly where Whisper operates. It gathers context across systems and prepares specific actions, which it then executes directly once approved by a human.
“It felt logical to start with customer support, which usually requires moving across most company systems to resolve an issue. It’s also where we have the most experience, and we don’t see an optimal solution on the market. We don’t want to be just another layer that produces nicer text. We want requests to be truly resolved — from start to finish — and for the customer support team to have a single place from which they can work.”
In customer support, writing the email itself isn’t what takes the most time — it’s everything around it. To resolve a single request, a person often needs to open five different windows: find the order, verify shipment status with the carrier, notify the warehouse, or update a shared spreadsheet. This “running around” between systems and actions within them is exactly where Whisper operates. It gathers context across systems and prepares specific actions, which it then executes directly once approved by a human.
Whisper works as a plug-and-play solution not only on standard platforms like Shoptet or Shopify, but on virtually any system with an API and reasonable documentation. It does so without expensive development or complex setup by engineers. It studies how a system works on its own so it can start operating in it immediately when prompted by a human.
“We’re targeting mid-sized e-commerce players who want efficiency, but where accuracy and control over changes are still critical. Whisper has to be able to handle the system fragmentation that’s typical in e-shops.”
Customer support is just the beginning. Over time, Whisper will assist in other e-commerce areas where context must be gathered and actions executed across systems — from returns and claims to communication with carriers or suppliers, and even automated product catalog management. Whisper is currently being tested by e-shops of various sizes, including billion-crown retailer Lentiamo.
“We have high expectations for Whisper. We were looking for a tool that would use AI to automate a large part of our customer support. Most tools can only summarize a customer request and suggest a reply — Whisper goes much further. Once the entire workflow is fine-tuned, it should be possible to process most requests fully automatically."
