Editor's note
AI is steadily moving out of dashboards and into daily field work.
This week’s stories show a clear pattern: AI assistants accessible from a phone, real-time machine data streamed from harvesters, log measurement done on-site with a smartphone, and digital twins updated continuously. The common denominator is proximity. The tools are no longer abstract systems sitting in an office, they are becoming operational companions.
What matters is not that AI exists, but that it is reachable in the moment decisions are made. That shift alone has the potential to improve efficiency across the value chain.
Axel
WHAT GOT ME THINKING
UPM Scales Aarnibot AI Assistant to Hundreds of Finnish Foresters
UPM's generative AI assistant Aarnibot is now being used by hundreds of foresters in Finland, with plans to roll out similar AI work coaches across UPM operations in other countries throughout 2026. The AI provides instant answers to field forest experts on topics like FSC certification requirements and EUDR compliance, turning complex regulations into accessible field guidance. It's a practical example of how large forest companies are deploying LLMs to support operational decision-making rather than replace human expertise.
Axel’s notes: AI agents are no longer experiments confined to innovation teams. Aarnibot is being used directly by field foresters to interpret certification rules and compliance requirements in real time. That matters because regulatory complexity rarely slows down—but the time available to interpret it does.
This mirrors a broader trend we’ve touched on before: AI becoming a field-level assistant rather than a back-office analyst. When guidance, calculations, and documentation support are available directly from a phone, the efficiency gains compound.
The future of operational AI may not be about larger models, but about smarter, context-aware agents embedded exactly where decisions happen.
New Zealand Pilots Real-Time Machine Data to Boost Forest Productivity
New Zealand's FGR Automation & Robotics programme has developed a system that taps into harvesting machinery's CAN bus to extract operational data in real-time via Starlink. The cloud-based dashboard tracks operating time, fuel consumption, and terrain conditions, while machine learning predicts forwarder activity—loading, unloading, or traveling—with over 95% accuracy. The system can even analyze joystick movements to identify operator efficiency patterns, enabling standardized training and self-improvement through accurate feedback.
ForestX Acquires LogStack Timber Measurement Technology
Announced February 10, Sweden-based ForestX has acquired full ownership of the LogStack suite of timber measurement products. LogStack includes LogStackPRO (smartphone-based log pile measurement) and LogStackLIDAR (GPS-integrated scanning for comprehensive volume and diameter analysis). The products are currently used in Denmark, Germany, Hungary, and Poland, measuring over 2.5 million cubic meters of timber annually. The acquisition expands ForestX's digital product portfolio beyond consulting, giving them proprietary technology for the European sawmill industry.
Horse Powertrain Joins Scania's Electric Timber Truck Pilot in Sweden
Horse Powertrain has been selected by Scania to provide its 120-kW range-extender system for a heavy-duty electric timber truck pilot operating under SCA in northern Sweden. The hybrid solution combines battery-electric power with a compact multi-fuel generator, allowing the truck to complete 7–8 daily rounds (16 km each) through steep, remote terrain without charging downtime. A previous 100-day German trial achieved 90% CO₂ reduction compared to diesel, with over 90% of driving on pure electric power
GAR and Arkadiah Deploy AI-Powered Forest Carbon Monitoring in Indonesia
Announced at the Singapore Space Summit, Golden Agri-Resources (GAR) and climate tech firm Arkadiah are launching a 5-year project to deploy advanced digital MRV technology in West Kalimantan, Indonesia. The system combines high-resolution satellite data, LiDAR scanning (both aerial and ground), and AI-enabled geospatial modeling to create 3D digital twins of forest areas, tracking carbon sequestration over time. The partnership aims to overcome manual measurement limitations and establish scientifically rigorous baselines for tropical forest carbon stocks.