Microsoft AI Agents Adapted A Citrus Pest Tracker
Microsoft has published an engineering post describing how AI agents helped adapt an existing insect monitoring platform to track the Asian citrus psyllid, the insect that spreads the pathogen behind Citrus Greening Disease. The post went up on the Microsoft Discovery Blog on October 22, 2026.
Transcript
A tiny insect spreads a disease that threatens citrus trees. Microsoft says AI agents adapted its monitoring system.
Microsoft recently reprogrammed its insect monitoring stations to track the Asian citrus psyllid. The work ran through AI agents in its Discovery app.
The United Nations food agency estimates pests and plant diseases destroy up to forty percent of crop production each year. Citrus Greening Disease has no cure.
Microsoft says a literature review and experimental design normally take several weeks. The team compressed that into a few days.
The team has no permits for live insects, so the classifier trained on simulated wingbeats. No accuracy result is reported.
Inboxsmith builds AI phone and email agents for small business. Please like and subscribe for more news.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- Every narrated claim maps to a sentence in the first-party post. HOOK, insect spreads a disease threatening citrus trees: the post states the Asian citrus psyllid transmits an incurable plant pathogen to susceptible citrus trees that results in Citrus Greening Disease, called one of the biggest threats to the global citrus industry. HOOK, AI agents adapted the monitoring system: the post states the team used AI through the Microsoft Discovery app to explore how to adapt its platform from monitoring transmitters of animal and human pathogens to monitoring the Asian citrus psyllid. SCENE 1, reprogrammed insect monitoring stations to track the psyllid, and the work ran through agents in the Discovery app: the post says the Discovery app accelerated the process by which the team adapted its existing Biological Weather Station platform to lure the psyllid, and separately says that because these agents are also engineers they could directly reprogram the stations to support these experiments. SCENE 2, up to forty percent of crop production destroyed each year: the post states that the Food and Agriculture Organization of the United Nations estimates pests and plant diseases destroy up to forty percent of global crop production annually. SCENE 2, the disease has no cure: the same sentence describing the psyllid says it transmits an incurable plant pathogen resulting in Citrus Greening Disease. SCENE 3, a literature review and experimental design would normally take several weeks and were compressed into a few days: the post states that creating a comprehensive biomonitoring plan would typically require several weeks of research and planning, and that by enhancing the workplan with grounded agents the effort to perform a literature review, develop an experimental design, review biocontainment permitting requirements, and develop synthetic classifiers was compressed into a few days. SCENE 4, no permits for live insects and the classifier trained on simulated wingbeat data: the post states the app generated simulated psyllid wingbeats for testing in software prior to acquiring permits to work with live insects, and that the synthetic training data helped build an initial edge classifier that could be tested in the labs once permitted to receive live specimens. SCENE 4, no accuracy result is reported: the post reports only a preliminary dataset built to assess the false positive rate of the classifier against other insect species, and gives no detection accuracy figure anywhere in the text. The post is Microsoft writing about its own Discovery app and its own Microsoft Premonition program, which the narration frames by attributing each claim to Microsoft.(Microsoft Discovery Blog post by entomologist isaiahhoyer with co-authors Mike Reddy and Ethan Jackson, titled Adapting Biological Monitoring for Food Security with Agents, published on the Microsoft Tech Community on October 22, 2026.)