Designing AI chat experiences and multi-product integrations to support airplane maintenance communication and task-on-time.

ROLE

Product Design Intern

TIMELINE

Jun - Sept 2025

TEAM

Charlie Barr

Non Pronanun, Kush Goel, Cody Tan, William Eng

Jill Reed

Non Pronanun, Kush Goel, Cody Tan, William Eng

ORG

Maintenance & Engineering

BRIEF

Over my 12 week internship, I worked on 4 core products, shipped an end-to-end engineering product, and collaborated with product leadership to help shape future product vision.

I designed internal tools for thousands of aircraft engineers, users who need software that work fast, without disrupting their mental models.

I also got to zoom out. I partnered with design and product leadership on a vision workshop and put together a proposal for where the product org could go next. That was a new way of system thinking for me — thinking not just about what to build, but why, and in what order.

Project 01

iPadOS - Shipped

MxHub — AI Summarization Chatbot

Problem

Every flight sends a stream of coded aircraft data called ACARS, which take very long to interpret…

Technicians on the ground need to manually cross-check multiple apps before taking any action to resolve an ACARS message, leading to slower task completion and flight delays.

Challenge

Apply AI in a high-risk environment without making engineers dependent on it.

In aviation, safety always come first. The goal was to make AI visible enough to support faster decision-making, while keeping the engineer's judgment at the center.

Solution

Non-intrusive inline AI summary that appears at the top of the page.

Technicians get the key insight in seconds instead of minutes, helping them make faster, more confident maintenance calls and turning around aircraft with fewer delays, with no new prompting or onboarding training.

Users can select from guided prompts to ask follow up questions.

The guided prompts limit users and what they can ask so the AI doesn't give opinions or suggestions. It simply surfaces the facts, helping users make their own call.

Process

Ideation

With safety being prioritized, I had to introduce AI to users’ workflow without it being disruptive, intrusive, or dominant.

I brainstormed and analyzed AI chat tools such as ChatGPT, Apple Intelligence, and Gemini, to evaluate different entry points and chatting experiences that would best serve my users and safety. I asked myself: what makes AI supportive while secondary?