Context aware AI-powered video shooting copilot
ABOUT ▶︎
My Role Lead Product Designer
Team 1 Designer/5 Engineers/ 2 AI scientists/ 2 UXR
Outcome Product sold 2M globally, 25% higher MAU
DJI ShotGuide
WHAT I DID ▶︎
◆ Owner of product vision and strategy
◆ Interaction and visual design
◆ Led XFN collaboration
◆ Drove UXR and design workshops
◆ End-to-end product design from concept to launch
INTRO
User Problem
Most DJI gimbal users expect a much simpler vlogging experience, but often face a steep learning curve, resulting in reduced user satisfaction.
Glance of Solution
ShotGuide recognizes context and shooting targets, providing automated camera movement and assistance as a video shooting copilot.
IMPACT MEASUREMENT & PROBLEM RESEARCH
Data Analysis
To identify the impact of the user problem and define the measurement of success.
User Interview
Participant: 50+ beginner users in three counties 🇬🇧 🇨🇳 🇯🇵
Findings ▶︎
◆ 70% of users needed a lot learning to on board.
◆ MAU of beginner users had a 35% decline in 6 months after launching.
◆ 65% of participants stopped using because “hard to learn compositions and camera movement”
Findings ▶︎
Problem 1 Confusion around composition and camera movement
“I always find my video doesn’t reproduce my feeling to that environment when I was there”
Problem 2 Not motivated to learn more about videography
"This shouldn't be more complex than shooting video with my phone."
REFRAME PROBLEM AFTER RESEARCH
Beginner users struggle with selecting appropriate camera movements and compositions for their shooting target.
WORKSHOP
I hosted rounds of workshops with engineers, UXR, PMs. By introducing the design sprint process, I led weekly iterations and helped team to land on two concepts within two weeks.
DIRECTION EXPLORATION
Design sprint led to with 2 directions. I worked with UXR to invite 14 users to for concept test with the prototypes I built, and their reflections and opinions were analysed.
▶︎ Direction 1
Scenario-based tutorial and learning resources
Provide easy-to-follow video instructions for camera operator movement and gimbal holding technique, ensuring that users can reproduce high-quality videos with ease and zero friction.
▶︎ Direction 2 ▶︎ Final direction
A context aware AI-powered video shooting assistant
System recognizes the context and shooting target, and proactively provide shooting assistant, automated composition and camera movement.
WHY DIRECTION 2
I arrived at the decision to select direction 2: a context-aware AI-powered video shooting copilot as the final design direction. This choice was driven by three key factors:
Desirability
AI Co-creation experience give beginners superpower
Feasibility
DJI has existing automated camera movement research efforts.
Strategy
Push the company strategy of context understanding and co-creation
I designed a prototype to demonstrate a hero user flow and explore interaction design, which was also utilized for a user study to obtain feedback:
◆ Camera detects the environment and shooting target via live preview
◆ AI suggests appropriate shooting techniques with a simple footage.
▶︎ 1 Context awareness and AI suggestion
◆ User pressing the shutter button, the gimbal automatically tracks and adjusts the target's composition while conducting camera movements.
◆ The user can then compare the simplified video clips and decide whether to save the footage.
▶︎ 2 Video shotting copoilt
FINAL DESIGN CONCEPT
Iterated Design
PROTOTYPE & ITERATIONS
Adding Detailed Video Shooting Guide
User feedback:
Only showing the final result was not enough for them to learn how to hold the gimbal and move in the space.
Iteration:
Provide easy-to-follow video instructions for camera operator movement and gimbal holding technique,.
.png)
Scenario-based ShotGuide
User feedback:
After getting some cool footages, they don’t know what to shot nexts and need inspiration.
Iteration:
ShotGuide can also offer shooting guides based on scenario recognition, providing users with related shooting inspirations for future shoots.
▶︎ Context awareness and AI suggestion
FINAL DESIGN
While framing, the camera detects the environment and identifies potential shooting targets, and a pop-up modal suggests appropriate shooting techniques. Users can review sample footage and confirm or reject suggestions.
▶︎ Video shotting copoilt
UI guides user to hold the gimbal and move in the scene. The gimbal automatically tracks and adjusts the target's composition while filming.
▶︎ Scenario-based ShotGuide Library
ShotGuide library can be accessed in the main screen, providing users with related shooting inspirations for future shoots.
WAHT I LEARNT
User prototype as a medium for team alignment
Prototypes and demos can provide tangible representations of abstract ideas. In innovative projects where there are few similar examples to draw from, using demos not only facilitates user testing but also promotes better communication among team members, sparking discussions and collaboration.
Impact is not only about the user
The design decision is a balance between users, business, resources, and the stage of the product. When aligning with leadership to secure resources, focusing on the user problem and user impact is not enough. What does this project bring to the overall/future picture of the business and technology roadmap? And what does it cost to achieve the success matrix? Those impacts are as significant as the users' impacts when proposing the project to leadership.
Use design to inform tech research and investment
The relationship between design and technology goes beyond feasibility, as design can influence technological research and investment. In this project, we identified human-machine co-creation can be a great direction for companies future roadmap. Design envisioning can shape a more desirable future and provide insight into a company's future investment and technology roadmap direction.