An AI Virtual Assistant for Faster, Smarter Aircraft Maintenance
AVA, the Aero Star Virtual Assistant, surfaces aircraft maintenance history from Corridor so technicians can prepare, diagnose, and resolve work orders faster, with plain-language questions and cited answers.
Maintenance history, answered in seconds on the shop floor
Aero Star Aviation partnered with HatchWorks AI to design and deliver AVA (Aero Star Virtual Assistant), an AI platform that surfaces aircraft maintenance history from Corridor so technicians can prepare, diagnose, and resolve work orders faster.
Phase one targets technicians. Future phases extend to planners, managers, and customer service, and build a proprietary knowledge base.
The goal: equip technicians with instant, trustworthy answers from past work to reduce diagnosis time, raise first-time fix rates, and cut repeat repairs.
The answers existed. Nobody could find them fast enough.
Historical maintenance data in Corridor was unstructured and inconsistent, limiting reuse. Technicians couldn't quickly access prior fixes for similar issues, while ongoing mechanic shortages increased pressure on turnaround times.
What Aero Star was up against
-
01
Unstructured, inconsistent history
Maintenance records in Corridor were hard to reuse, so prior fixes for similar issues stayed buried.
-
02
Mechanic shortages, tighter turnarounds
Fewer technicians and more pressure on turnaround times left no slack for slow diagnosis.
-
03
Tribal knowledge, repeated mistakes
Critical know-how wasn't captured centrally, leading to repeated mistakes across work orders.
Map the workflow, clean the data, then build the assistant
-
01
Discovery and workflow mapping
Mapped technician journeys and common questions to desired responses and supporting data.
-
02
Data pipeline and guardrails
Orchestrated data pipelines in Airflow using DAGs to extract and normalize Corridor data for search and analytics.
-
03
MVP development
Delivered a secure, device-friendly AI assistant (desktop, tablet, mobile) on GCP with role-based access.
-
04
Validation loops
Iterated with stakeholders to tune accuracy, recovery flows, and UX for shop-floor speed.
Corridor history is normalized once, then every question is answered two ways: cited passages from past work orders, and structured facts pulled by generated SQL.
A production-ready assistant that meets technicians where they work
HatchWorks AI delivered a production-ready AI assistant that uses retrieval-augmented generation to answer technician questions with cited history. It adds text-to-SQL to pull structured facts from Cloud SQL, integrates Corridor via automated ETL and synchronization, and is deployed with Terraform-based infrastructure, centralized logging, and lightweight MLOps for smooth, reliable operations.
The solution meets technicians where they work, with plain-language questions and cited answers, leveraging Corridor while building an independent knowledge base. It runs on a cloud-native architecture that's easy to extend across roles and sites.
Faster diagnosis
Relevant history surfaces in seconds, cutting search time.
Higher first-time fix rates
Prior resolutions highlight pitfalls and proven steps.
Shared knowledge
Technician insights become searchable, reducing repeats and rework.
Scalable foundation
Ready to add pricing data and expand to planners, managers, and support.
Maintenance history in Corridor unstructured and inconsistent, limiting reuse
Technicians unable to quickly access prior fixes for similar issues
Tribal knowledge not captured centrally, leading to repeated mistakes
Corridor data extracted and normalized through Airflow pipelines for search and analytics
Plain-language questions answered in seconds with cited history and text-to-SQL facts
Technician insights searchable in an independent knowledge base, protected inside Aero Star's walls
Early signals from the pilot
drop in recurring actuator issues after AVA insights (pilot example)
NPS target from technicians using AVA
Shorter AOG turnaround time through faster troubleshooting
Cited answers from past work orders, plus structured facts via text-to-SQL on Cloud SQL
The 30% figure is a pilot example and the 85%+ NPS is a target, both as stated by Aero Star Aviation.
The stack
In their words
“HatchWorks AI didn't just drop in a chatbot. They took the time to map our workflows, clean up how we use our data, and build an assistant our team actually trusts. AVA captures our knowledge, protects it inside our walls, and gives us a platform we can keep building on for years.”
About the work
About Aero Star Aviation
Aero Star Aviation was founded in 2013, specializing in Embraer Phenom and Praetor maintenance and management dating back to the first Phenom production nearly 16 years ago. Aero Star Aviation is rapidly growing into the number one alternative for Embraer Phenom and Praetor maintenance and management in the southern hemisphere. Its employees are completely specialized and educated on the Embraer aircraft to ensure the highest quality service. With locations in Dallas, TX and Ft. Lauderdale, FL, the company prides itself on customer connections and superior service.
HatchWorks AI turns AI into ROI by automating key business processes, transforming data, deploying intelligent agents, and shipping AI-powered products that deliver measurable results. Its proprietary Generative-Driven Development methodology is a repeatable path from idea to production, blending AI, agents, and engineering to ship faster with less risk.
Ready to put your operational history to work?
See how HatchWorks AI builds assistants that answer from your own data, with citations your teams can trust.