CASES

SMART FARM started with a simple product hypothesis: if cameras could be trained to recognize key situations on a dairy farm, video monitoring could become a real operational tool rather than passive surveillance.

Farm teams already rely on visual observation to notice feeding issues, health risks, and routine violations. But on large farms, these checks are manual, fragmented, and easy to miss.

The challenge was to turn raw camera footage and AI detections into clear workflows: what happened, where it happened, which animals or groups were affected, and what action should be taken.

My role was to design the product logic and interface that made these detections understandable and useful for everyday farm operations.

We were designing a new operational layer for dairy farms — a system that could turn camera footage and AI detections into decisions people could trust and act on.

The product logic had to be defined almost from scratch: what situations were worth detecting, how different farm roles interpret them, and what context they need before taking action.

A detection alone was not enough. Users needed to understand what happened, where it happened, which cow or group was affected, how serious it was, and whether the system’s conclusion could be trusted.

The interface had to make complex herd health, feeding, and milking routine data clear enough for everyday farm work.

At this stage, the biggest unknown was not the interface itself. The bigger question was how dairy farms actually make decisions and how AI detections could fit into daily operations.

I used research to understand the farm as an operating system: who notices problems, who is responsible for reacting, what data people trust, and what context they need before taking action.

I started by analyzing existing farm management systems, monitoring tools, and agricultural BI products to understand common workflows, reporting patterns, and information architecture.

Then I conducted field research on real dairy farms (with >800 lactating cows). I observed feeding routines, animal movement, communication between roles, and how decisions were made during the day.

The goal was not to create abstract personas. It was to define the product logic: which situations should become alerts, what level of detail users need, and how the system should connect farm-level analytics with group, cow, and operational workflows.

Farm teams don’t need another dashboard —
they need a connected view of what is happening.

Research showed that farm data was not simply missing — it was fragmented.

Health records, feeding data, milking results, camera footage, and operational notes often lived in separate systems or in people’s heads. This made it hard to understand whether a problem was isolated, recurring, or connected to a broader pattern.

This insight shaped the product architecture. Instead of designing separate tools for each module, we built SMART FARM as a connected platform that combines external data sources, computer vision detections, historical records, and workflows across individual cow, group, and herd levels.

Each farm can select and arrange widgets based on its own operational priorities.

Detected deviations are designed as alerts.

Each item gives enough context to understand the issue at a glance, while the detail view provides visual evidence, key metrics, and an activity log for follow-up.

Deviation detected → Notification sent → Review evidence → Decide what to do

Users can move from a structured herd list or group to a unified profile for any cow. Identification, group, lactation, milk production, health indicators, and transfer history are brought together in one place, reducing the need to switch between separate farm systems.

The platform supports more than 25 operational workflows — from animal identification and health inspections to feeding management, analytics, and access control.

Most farm work happens away from a desk. The mobile experience gives on-site teams quick access to the farm overview, current deviations, key metrics, and individual animal data.

Rather than reproducing the entire desktop platform, the mobile interface prioritizes the information users need while working in the barn.

The design system started with the very first module, not as a separate initiative. Every new module added only the components and patterns it genuinely needed, so the system grew alongside the product rather than being built upfront and imposed on it.

Dingo started as an international calling app and was expanding into a broader travel connectivity product with eSIM.

This created a new journey that did not end at purchase. To receive value, users still had to choose the right plan, understand device compatibility, install the eSIM, configure their phone, and confirm that the connection was ready to use.

My role was to redesign the product structure and create a clear path from choosing a plan to the first successful connection, while keeping international calling accessible within the same app.

The challenge was to redesign Dingo from a calling-only app into a broader mobile connectivity product.

Adding eSIM introduced a new user journey: choosing a destination, comparing plans, purchasing, installing, and activating the service.

The experience had to feel clear, trustworthy, and easy to complete without support — while keeping international calling accessible inside the new product structure.

Before redesigning the product, I analyzed the existing mobile app, its calling-focused structure, and the new requirements introduced by eSIM.

The goal was to understand which parts of the old experience could be reused, where the existing structure created limitations, and what new flows were needed to support travel connectivity.

I also reviewed common eSIM purchase and activation patterns to identify the moments where users need the most guidance: destination selection, plan comparison, payment, installation, activation status, and troubleshooting.

Users don’t want to manage telecom complexity —
they want to know that they will be connected when they need it

The most fragile part of the experience was not the purchase itself, but the uncertainty around it.

Users needed to understand whether the plan worked in their destination, what would happen after payment, how to install the eSIM, and whether the service was ready to use.

This shaped the product approach: the interface had to guide users step by step, explain important details before purchase, and make service status clear after activation.

I reworked the mobile app around user intent rather than the old calling-only structure.

The new structure separated the main needs: getting connected with eSIM, managing active services, checking balance, viewing call rates, and making international calls.

This allowed eSIM and calling to coexist in one product without competing for attention.

Choose destination → Select plan → Review purchase → Confirm

After purchase, the app guides users through installation and makes the eSIM status clear once the service is ready to use.

Calling remained part of Dingo, but was redesigned as one of several connectivity scenarios rather than the only main use case.

Senior Product Designer based in Prague
6+ years in digital product design

I lead design from discovery to delivery — turning research, business goals, and technical constraints into clear product systems and scalable user experiences.

My strongest experience is in 0-to-1 products, AI-powered platforms, and complex operational workflows. At SMART FARM, I built the design process from scratch and designed a web and mobile platform used across more than 25 farms.

I work closely with product managers, engineers, and domain experts to make sure design does not only look clear, but helps users make decisions, complete tasks faster, and trust the product.

Creativity has always been part of my life. Outside of work, I paint with oils, enjoy sewing and fashion, and love travelling and discovering new places.