Bluetail Eclipse
Redefining aircraft records keeping with
generative AI
I led the UX design for a system that reads PDFs, intelligently extracts relevant data, and creates organized, digital logbooks. Working closely with aviation experts and engineers, we crafted an experience that bridges advanced AI capabilities with intuitive workflows for maintenance teams.
My Role
Lead Product Designer
Team
Bluetail Product Trio
Tools
Figma, Miro, Jira
Timeline
6 months
The Problem
Unstructured logbook records and challenges with digitization make it difficult for Bluetail users to maintain records as part of daily operations.
Aviation maintenance teams face time-consuming manual work extracting and organizing data from countless aircraft PDFs. This error-prone process threatens regulatory compliance, increases costs, and causes stress for operators during audits.
Solution Highlights
Accelerated ingestion with AI for faster, consistent uploads.
We streamlined the upload process by adding AI-assisted ingestion, reducing manual work. Users can continue uploading records with minimal effort while getting standardized, accurate results.
Introduced a structured logbook view for reliable research and daily use.
I designed a dedicated logbook layout that moved beyond file storage. Users can now search entries by keyword or review logs front-to-back, giving them confidence to conduct research and use Bluetail as a living logbook.
Impact
The redesign moved Bluetail closer to FAA recognition of digital logbooks.
The Process
Our team works in an agile method paired with design thinking
We follow agile thinking method as a team and I followed a form of design thinking method
Our Team
Our Goal
To create a an MVP that helps the departments ingest files more effectively and access them in a structured, centralized way—replacing fragmented, manual workflows.
Understanding the Problem Space
I investigated how technicians, DOMs, and RMOs currently organized logbooks and their hesitations about digital records.
Digital Ethnography
I investigated how technicians, DOMs, and RMOs currently organized logbooks and their hesitations about digital records.
Our research surfaced three organization methods. We selected the highest-value pieces to build—those that advance traceability, audit-readiness, and Bluetail’s “source-of-truth” vision.
Secondary Research
Created physical prototypes by printing logbook pages and re-filing them according to user strategies to better understand real-world workflows.
Created physical prototypes by printing logbook pages and re-filing them according to user strategies to better understand real-world workflows.. These findings grounded the problem space and informed early design decisions.
Feature prioritization
Aligning Scope, Risks, and the Migration
As a team, we sequenced priorities, addressed engineering risks, and planned the transition from Bluetail’s file-centric model to page-based records.
Selected Features
Based on this prioritization, I moved forward with designing the experience for two features: Logbook front & back review and AI ingestion files.
Final Designs
Logbook Viewer
Use case 1: Front and back review of entries
As a Records Manager (Airframe), I want to complete a front-and-back review of each logbook entry so that every page is complete, credible, and audit-ready.
Use case 2: Add and View Supporting Documents
As a Records Manager, I want to view and add supporting documents to a logbook entry so that the evidence for that entry is complete, traceable, and audit-ready.
Use case 2: Edit Logbook
As a Records Manager/Technician, I want to edit logbook entries so that every record is accurate, complete, and audit-ready.
AI Ingestion
Use case 1: Front and back review of entries
As a Records Manager (Airframe), I want to complete a front-and-back review of each logbook entry so that every page is complete, credible, and audit-ready.
Reflection
Through research and usability tests, a few patterns stood out that now guide how Eclipse is designed and built.
This project challenged because I’d never shipped an AI workflow before, and Eclipse had almost no comparable products—both forced me to learn fast and invent carefully.
Learnings
Designing AI users can trust
Inventing patterns for a one-of-a-kind tool
My biggest stretch was shaping AI interactions for older, less-technical users. I studied established patterns, tested plain-language prompts, and made every suggestion reviewable and reversible with visible confidence.
With few benchmarks, I modeled the domain from the ground up—page as the atomic unit, relationships for evidence, and guided flows with optional flexibility. Rapid paper→prototype→test loops helped translate hangar workflows into UI without copying the wrong conventions.