As deployments of edge AI scale in the farming sector, continuous monitoring of edge fleets – literally in the field – becomes impractical. Autonomous machines create value when they operate without human oversight and request attention only when needed.
Machines like those from Burro move loads and travel between working areas in vineyards and farms. Their usefulness rests on their ability to move and operate inside software-defined boundaries, and to signal exceptions reliably.
Operators can’t track the movement of every machine, despite the best efforts of dashboard designers. Similarly impractical is watching a dozen or a hundred live video feeds, even when conditions allow such a set-up to work out in the open. Mechanisms are better designed to automatically filter all inputs and work instead of, and at a greater scale than a human operator’s attention.
A system built recently by Akamai and Agri Automation Australia monitors location data from the Burro Cloud API, evaluates it in the context of pre-defined geofenced areas, and issues notifications when one or more conditions are met. A robot entering a loading zone or storage facility, or moving close to a public access point will trigger events, such as an automated message.
The logic of the setup runs on Akamai Functions, the company’s serverless execution environment. Functions execute code that’s been compiled to WebAssembly. Code runs don’t persist beyond the duration of each invocation, so there’s no need for large-scale server provision to host thousands of lines of code. The function is invoked, a task is performed, and the code instance exits.
Each execution retrieves the latest robot position, checks it against geofencing rules, and decides whether a notification should be sent. Each state is persisted in managed storage so no duplicate notifications appear. The design ensures no long-running processes run that need monitoring, there are no scaling issues that would need expert systems administration, and there’s no dependency on a data centre and connection to it.
Akamai Functions operate inside a distributed edge platform built originally to handle web traffic. The properties that benefited high-scale web serving also work in agricultural settings, the company says. Latency is low due to execution occurring near the point of request, yet availability is high because the platform covers several locations. The WebAssembly runtime restricts access to the host environment, and code is transitory.
The company’s Functions platform is finding an increasing number of uses in the agricultural sector, an area, among others, it will be showcasing at the upcoming TechEx North America event (see link in article footer).
On farms and other agricultural settings, locations where the technology is deployed can be dispersed, with varying degrees of connectivity. Depending on the weather and time of year, the nature and scale of required workloads can change. In these contexts, a dependence on a central backend or constant network connection can create a meaningful level of error and fragility.
The nature of edge execution means the processing of events close to the data sources. A function may call a cloud API for location, for example, but as the decision logic runs at the edge, there’s a much shorter path between data retrieval and any needed physical intervention.
The fact that end-users are charged per-invocation and resulting compute time means much lower costs than those of pre-provisioned capacity – ideal for event driven workloads. Notification functions, for example, only trigger costs when they run, and there’s no ‘standing charge’ for idle resources.
Like all good technology, a modular, incremental solution can be built over time. Akamai Functions can be integrated with other services running on the platform, including traffic management, cache-ing, and enhanced cybersecurity. Geofencing logic can be altered without changing the deployment model, new notification methods can be added (perhaps dictated by existing farm management software’s methods). Systems are easily replicated on several sites with minimal changes, with core logic remaining much the same, and only location-specific configurations changing.
Navigation, perception, and control remain can remain on the smart agri-robot or device. In these instances, the edge function acts as an intermediary layer, interpreting output from each robot or its cloud interface, and determines whether to involve the human operator. Inference can continue to take place on-device, handling tasks like obstacle detection or path planning, enhanced by edge functions handling aggregation and policy enforcement. A model detecting an anomaly in crop conditions or equipment can let the edge platform decide whether it meets the threshold for escalation and notify an operator.
Clearly, the effectiveness of any system rests to a certain extent on the quality of location data and the definition of geofences. Connectivity between robots or machines, the cloud API, and the edge platform must be sufficiently reliable: While edge compute reduces latency, it doesn’t remove the need for reliable data.
Akamai Functions and similar stacks provide a way to implement the balance between edge, cloud, and automated worker without building and maintaining an infrastructure. Keeping it simple – to let farmers and agricultural workers concentrate on their tasks – means not introducing unnecessary complexity into any system designed to reduce labour and improve yields.
(Image source: “Male mechanical engineer with sustainable agricultural robot in field” by This is Engineering image library is licensed under CC BY-NC-ND 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-nd/2.0)
Want to learn how Akamai Technologies is applying edge computing, IoT, and AI in practice? As a track sponsor at Edge Computing Expo North America 2026, Akamai will be speaking on the Edge Computing & AIoT track on Day 1, with attendees able to hear directly from their team at the San Jose McEnery Convention Center on May 18-19, 2026.
IoT News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.