Oxolo: redesigning GenAI video editor for B2B
overview
services
Product Design, Research
role
I led the redesign closely collaborating with the CEO, CPO and engineers.
year
2023-2024
platform
Desktop web
challenges
After reviewing the existing product, customer requests, customer journey, and aligning with stakeholders, we decided to focus on the following challenges:
challenge #1
How might we help users produce quality videos as quickly as possible?
When starting to create a video, users had to choose one of four ways (generate video from URL, bulk creation, text to video, or use a template). Starting a video with displaying all the possible options at once led to extra effort from the user to decide what option suits them best.
AFTER
After interviewing users, it became clear that most of them create videos from templates (including their own). The new video creation flow now starts from selecting a template: either from their own or provided by Oxolo. Customers can also skip the template and start their video from blank.
Create new video starts with selecting a template.
TO PROMPT OR TO generate from url?
Once template is selected, users can generate a video either from a link (for an existing product) or from a prompt (for an upcoming product or any topic).
When generating video from a link, Oxolo scrapes the website to get images and copy as input for video. If there's no link to create a video from, user can make a video from a prompt and add images themselves.
"
I really like starting from a template as it will save me time and work. Currently, I use ChatGPT for generating script and Canva for visuals, so starting with a template will help me save time on having to use both.”
Kimberley, D2C brand owner
"
I used to duplicate my previous videos when creating a new video, so being able to start with my own template is great.”
Peter, luxury brand owner
challenge #2
How might we make the video editor easy to use for customers without video making experience?
research
After examining usage analytics and watching users interact with the product live through Hotjar, it became clear that once video is generated, the main actions users wanted to take were:
1
Edit generated video script
2
Change AI avatar and voice
3
Customize branding
action #1
Edit video script
Once a video was generated, users had to switch between scenes/slides to see the AI generated script. There was no overview of the whole script and it was hard to tell if the story holds up throughout the video.
AFTER
Introducing a separate Script panel enabled users to easily edit the entire video script, change avatar, language, and voice—the most common actions that users were performing.
action #2
Change AI avatar
and voice
Before, users had to spend extra time and effort to update avatars and voices since they were separated.
AFTER
Now, users can easily update avatars and voices in their videos, to see what works best for them using Script or Properties panel . Users can choose between applying the same voice and actor to the whole video or to variate them scene by scene for more engagement.
action #3
Customize
branding
In the previous version, branding capabilities were limited by media. Users couldn't save other important branding components like colors, fonts, and templates to use across videos.
AFTER
Users can make generated videos fit their brand effortlessly with Brand Library. There, they can customize and reuse images, clips, colors, fonts, custom voices, avatars, and templates across different videos.
Over 6 months, I helped design and ship entire GenAI video platform, including:
"At Oxolo, Fedor single-handedly helped our startup mature by elevating our UX and product design in just weeks. He consistently demonstrated exceptional speed, responsiveness, and thoughtfulness. Fedor's impact was felt across the entire product, and his expertise is invaluable to any team."
Victoria Busse,
CPO at Oxolo (formerly TIER, eBay, Blinkist)
