Powered by a new agentic AI Adaptive Clustering engine, grounded in over a decade of curated maintenance outcomes - built to get sharper with every fleet.

Veryon, the leading provider of aviation maintenance software and data intelligence solutions, today announced the next generation of its Defect Analysis platform, now expanded to incorporate aircraft faults and part failures, directly into every chronic - accelerating an investigation process that has historically taken engineering teams weeks to cross-reference data sources and confirm repeat-defect candidates before a task card can be issued.

The announcement comes as operators face mounting pressure to move from reactive to proactive maintenance, eliminating chronic defects, which drive delays, cancellations, and unnecessary parts and labor costs. Veryon's next generation platform is designed to compress what used to be a lengthy, root-cause analysis into same-day, AI-assisted detection - flagging a chronic issue, and the part driving it, before it grounds another tail.

Built Against a Verified Benchmark, Not Just a Model

Most AI in aviation maintenance today is pattern-matching on raw data with no ground truth behind it. Veryon takes a different approach: rather than tuning its clustering algorithm against isolated operator provided assumptions, Veryon updated its “golden dataset” incorporating chronic defects - independently reviewed and verified by aviation maintenance experts - and used it as a standing benchmark for every AI model change. That updated benchmark work, now coined, Veryon’s Verified Data Set, has driven accuracy improvements to be the highest in the industry.

"Most vendors show you an accuracy number and ask you to trust it. We work with our customers to demonstrate results that they can trust, incorporating their operational data into a benchmark,” said Vinay Kumar, CTO, Veryon. "That's the difference between an AI system getting more aggressive and hallucinating vs an AI powered system getting more correct - and in a regulated environment, that difference is the whole point."

A Retraining Loop the Maintenance Desk Controls

Veryon’s Verified Data Set gives every operator a common, expert-reviewed foundation for detecting chronic defects. On top of that foundation, an agentic AI engine learns from each operator's own maintenance team - the combination is what Veryon calls Adaptive Clustering - a proprietary method that tunes defect-detection accuracy to each operator's specific fleet and operating history.

The system stays under human control at every step. Every AI-generated suggestion passes through review before any action is taken. When an engineer approves, splits, or merges a chronic, that decision doesn't just update a single grouping; it feeds a retraining loop that adapts the clustering logic for that operator's fleet going forward. Before any of those updates take effect, the operator's own maintenance desk reviews them, so the AI's judgment evolves under the same review discipline that already governs the maintenance record.

A Platform for Maintenance & Reliability Teams

Veryon's next generation Defect Analysis includes new maintenance insights, alongside an expanded interface and a set of AI-powered tools:

• Chronic Detection powered by AIRE Adaptive Clustering — Veryon’s Verified Data Set paired with an agentic engine that learns from engineers' approve/split/merge decisions, tuned to each aircraft make and model, and gated through operator review before any change goes live.
• Veryon AIRE Assist — a natural-language assistant that lets maintenance teams investigate fleet defects conversationally, grounded entirely in that operator's own defect and chronic history.
• Fault Data Integration — embeds fault data directly into the defect lifecycle, closing the fix-confirmation loop without the 30-to-45-day wait maintenance teams previously faced.
• AI-Generated Chronic Summaries — a single narrative per chronic issue, synthesizing logbook entries, sensor faults, parts history, and utilization trends into one readable summary.
• Rogue-Parts Detection — flags components statistically linked to recurring chronic defects across the fleet, so operators can act on the part driving a chronic, not just the chronic itself.
• Fix Agent powered by Fleetwide Data Intelligence — a conversational fix-recommendation layer drawing on field-tested solutions from decades of maintenance outcomes across commercial and business aviation fleets.

About Veryon Defect Analysis

Veryon Defect Analysis is built on more than a decade of confirmed, human-reviewed maintenance outcomes, drawn from roughly 30% of the world's fleet - a dataset no OEM or single operator holds on its own. Clustering accuracy is measured continuously against Veryon’s Verified Data Set, combined with the operator’s uniquely trained agentic algorithms trained to independently verify chronic defects for their fleet, with every model change tested against that benchmark before release.

About Veryon

Veryon is the leading provider of aviation software and information services, trusted by nearly 6,000 customers, 75,000 maintenance professionals, and more than 100 OEMs in nearly 150 countries. Powered by the world's largest de-identified aviation maintenance dataset, the all-in-one Veryon suite combines OEM-authorized publications, intuitive cloud-based maintenance workflows, and AI driven insights to help operators, MROs, and OEMs maximize flying time without compromising safety or compliance.

Veryon's Fleet Performance Partnership program, brings together Veryon's platform capability with expert guidance, structured training, and ongoing support to help operators realize faster adoption and stronger operational outcomes. Drawing on Veryon's 50+ years of experience and support from experts who understand aviation, Veryon customers have achieved up to a 75 percent reduction in troubleshooting time for new technicians and up to a 23 percent reduction in downtime costs. Learn more at veryon.com.

View source version on businesswire.com: https://www.businesswire.com/news/home/20260909237744/en/