Case Study: Bee friendly
Overview
Bee friendly builds autonomous sensor stations designed for real-world, remote monitoring of insects and biodiversity. Their devices—deployed both in fields and at enterprise client sites—collect data on insect activity and local climate using robust sensors, on-device AI, and cloud integration. At the heart of each system is a Raspberry Pi Zero 2 W with a Luxonis camera providing edge AI capabilities, all powered by solar and battery, and connected over energy-efficient wide-area networks such as NB-IoT.
Table of Contents
Challenge
Moving from prototype to product is never simple—especially when devices must survive off-grid and integrate seamlessly into enterprise infrastructure. Bee friendly needed a way to manage scheduled downtimes, optimize for solar energy, and stay reliably connected, whether via NB-IoT or other emerging technologies. Managing updates and device health at scale, while meeting customer- and region-specific requirements, added another layer of complexity.
What’s unique about an AI-based sensing product is that it’s never truly “done.” Algorithms are constantly improved, models retrained, and data pipelines updated. This is fundamentally different from traditional engineering projects, where hardware and firmware often remain static once shipped.
Bee friendly also anticipated that both connectivity and hardware needs would shift over time. As new base boards, camera types, or even 5G networking become attractive, their deployment stack needed to support easy transitions—without costly rewrites or field visits.
One of the most powerful enablers for Bee friendly’s engineering team is that make87 allows direct, remote development on live client deployments. If a problem is reported—or if a model needs to be debugged—they can set breakpoints, inspect system state, and even live-patch code on devices deployed in the field or at client sites. This dramatically reduces support friction and ensures issues can be resolved rapidly, without waiting for physical access or elaborate onsite setups. For a data-driven AI product that is continuously evolving, this ability to debug, improve, and test in real production environments is a step-change over legacy IoT solutions.
Results: Rapid, Reliable Deployment and Effortless Scale
With make87, Bee friendly moved from prototype to field-ready product in less than two months. Engineers focused on advancing AI, power management, and the core product—without reinventing infrastructure. The team now supports customer deployments around the world, including at enterprise sites where automated, high-integrity biodiversity data helps organizations fulfill EU reporting obligations (such as CSRD/ESRS E4 compliance). Reporting workflows are fully integrated, translating raw sensor data into standardized, legally relevant outputs for client documentation.
This chart shows one of the many metrics automatically tracked and included in client reports.
As the customer base grows and needs shift, Bee friendly can integrate new hardware, update AI models, and adapt connectivity strategies without friction. The result is a living, improving product—never static—delivering up-to-date, actionable insights to every client.
Technical Takeaways
make87 gives Bee friendly a reliable, scalable way to manage distributed, solar-powered AI devices:
- Raspberry Pi Zero 2 W and Luxonis camera with on-board AI handle all logic and sensing
- Devices connect using NB-IoT today, but can move to 5G or other networks in the future
- Power management and OTA updates are orchestrated from the cloud
- Engineers deploy, monitor, and upgrade the whole fleet with no manual intervention
- Automatic generation of CSRD/ESRS E4-compliant biodiversity reports
Conclusion
Bee friendly’s journey with make87 shows how flexible infrastructure empowers rapid product evolution—enabling a continuous loop of improvement, adaptation, and scale. By focusing on application logic and customer outcomes, the Bee friendly team delivers a modern, living product that meets both technical and regulatory demands across diverse deployment environments.
