Redesigning PixelBin as an AI generation platform
Role: Lead Product Designer | Duration: 6 months | Team: 1 Designer, 2 Engineers, 1 PM
Redesigning PixelBin as an AI generation platform
Role: Lead Product Designer | Duration: 6 months | Team: 1 Designer, 2 Engineers, 1 PM
Redesigning PixelBin as an AI generation platform
Role: Lead Product Designer | Duration: 6 months | Team: 1 Designer, 2 Engineers, 1 PM
Redesigning PixelBin as an AI generation platform
Role: Lead Product Designer | Duration: 6 months | Team: 1 Designer, 2 Engineers, 1 PM
01
Problem: a platform at a turning point
Problem: a platform at a turning point
Problem: a platform at a turning point
PixelBin launched as a toolkit for editing images at scale. With tools like watermark removal, upscaling and background erase, it let users do in seconds what previously required Photoshop skills.
PixelBin launched as a toolkit for editing images at scale. With tools like watermark removal, upscaling and background erase, it let users do in seconds what previously required Photoshop skills.
PixelBin launched as a toolkit for editing images at scale. With tools like watermark removal, upscaling and background erase, it let users do in seconds what previously required Photoshop skills.
But by 2025, the product was at an inflection point. Generative AI had fundamentally shifted user expectations. People no longer wanted to edit what they had, they wanted to create what they imagined.
But by 2025, the product was at an inflection point. Generative AI had fundamentally shifted user expectations. People no longer wanted to edit what they had, they wanted to create what they imagined.
But by 2025, the product was at an inflection point. Generative AI had fundamentally shifted user expectations. People no longer wanted to edit what they had, they wanted to create what they imagined.
But by 2025, the product was at an inflection point. Generative AI had fundamentally shifted user expectations.
But Pixelbin, which was focused was on editing and specific tools, rather than generation hadn’t caught up to the new user expectations.
But Pixelbin, which was focused was on editing and specific tools, rather than generation hadn’t caught up to the new user expectations.
But Pixelbin, which was focused was on editing and specific tools, rather than generation hadn’t caught up to the new user expectations.
But Pixelbin, which was focused was on editing and specific tools, rather than generation hadn’t caught up to the new user expectations.
Pixelbin dashboard before redesign
03.3
Who we designed for
Who we designed for
Marketing leads
Marketing leads
Needs fresh campaign creative fast, without briefing a designer or photographer each time
Needs fresh campaign creative fast, without briefing a designer or photographer each time
New to AI generation tools, so expects something closer to typing an idea than operating software
New to AI generation tools, so expects something closer to typing an idea than operating software
Judges the product on output quality and speed, not on feature depth
Judges the product on output quality and speed, not on feature depth
WHAT THIS MEANT FOR PIXELBIN
WHAT THIS MEANT FOR PIXELBIN
Navigation had to lead with "what can I make" rather than "which tool do I open," and menu labels had to stay in plain, non-technical language.
Navigation had to lead with "what can I make" rather than "which tool do I open," and menu labels had to stay in plain, non-technical language.
Content managers
Content managers
Handles high-volume production, needs the same edit applied across hundreds of SKUs
Handles high-volume production, needs the same edit applied across hundreds of SKUs
Came to PixelBin for specific tools like background removal, upscaling, not generation
Came to PixelBin for specific tools like background removal, upscaling, not generation
Thinks in outputs and batches, not prompts or models
Thinks in outputs and batches, not prompts or models
WHAT THIS MEANT FOR PIXELBIN
WHAT THIS MEANT FOR PIXELBIN
Batch Editor had to absorb AI features without losing the bulk-processing workflow this user already relied on
Batch Editor had to absorb AI features without losing the bulk-processing workflow this user already relied on
Agencies & Designers
Agencies & Designers
Juggles several client accounts and needs to switch context quickly
Juggles several client accounts and needs to switch context quickly
Already fluent in creative tools, wants fine-grained control over models and parameters
Already fluent in creative tools, wants fine-grained control over models and parameters
Evaluates PixelBin against dedicated AI generation tools, not just other editing suites
Evaluates PixelBin against dedicated AI generation tools, not just other editing suites
WHAT THIS MEANT FOR PIXELBIN
WHAT THIS MEANT FOR PIXELBIN
The generation studios needed advanced controls (model choice, duration, aspect ratio) surfaced for power users, not hidden behind simplified defaults
The generation studios needed advanced controls (model choice, duration, aspect ratio) surfaced for power users, not hidden behind simplified defaults
03.3
My role
My role
Rethink
Rethink all the navigation and architecture to position AI generation as the platform’s primary capability
Rethink
Rethink all the navigation and architecture to position AI generation as the platform’s primary capability
Redesign
Redesign the existing studio environments to cater to image and video generation and fit them in the existing console
Redesign
Redesign the existing studio environments to cater to image and video generation and fit them in the existing console
Rebuild
Rebuild the Batch Editor, the platform’s primary power-user editing environment, to include AI features
Rebuild
Rebuild the Batch Editor, the platform’s primary power-user editing environment, to include AI features
03
Solutions
Solutions
03.1
Old Signup
Only talks about B2B features with a text heavy styling
Old Signup
Only talks about B2B features with a text heavy styling
Old Signup
Only talks about B2B features with a text heavy styling
New Signup
New Signup
New Signup
Showcasing the new features with a refreshed UI styling
Showcasing the new features with a refreshed UI styling
03.2
New navigation and dashboard
New navigation and dashboard
New navigation and dashboard
Navigation based on the type of output the user needs as well as a dashboard better catered to AI powered environments.
Navigation based on the type of output the user needs as well as a dashboard better catered to AI powered environments.
Move navigation to top nav and make the menu items more layman friendly
Move navigation to top nav and make the menu items more layman friendly
Move navigation to top nav and make the menu items more layman friendly
Main section to now show different kinds of generation and editing environments
Main section to now show different kinds of generation and editing environments
Main section to now show different kinds of generation and editing environments
Add spaces to announce new releases
Add spaces to announce new releases
Add spaces to announce new releases
Keep a small tool focused section for legacy users
Keep a small tool focused section for legacy users
Keep a small tool focused section for legacy users
Show users various different types of usecases that Pixelbin can cater to
Show users various different types of usecases that Pixelbin can cater to
Show users various different types of usecases that Pixelbin can cater to
Show users various outputs they can get with just one input image- from images to videos
Show users various outputs they can get with just one input image- from images to videos
Show users various outputs they can get with just one input image- from images to videos
Show users generations made by other Pixelbin users to build trust factor
Show users generations made by other Pixelbin users to build trust factor
Show users generations made by other Pixelbin users to build trust factor
Maintain developer tools at the bottom for legacy users
Maintain developer tools at the bottom for legacy users
Maintain developer tools at the bottom for legacy users
03.3
New editing and creation environments
New editing and creation environments
New editing and creation environments
Updated with all the latest generation models as well as fine tuning controls for highest creative output
Image Generator
Generate original images from a text prompt. Choose from multiple AI models, iterate, refine, and export your genrations, all without leaving the studio.
Image Generator
Generate original images from a text prompt. Choose from multiple AI models, iterate, refine, and export your genrations, all without leaving the studio.
Video Generator
Turn text or images into videos using the latest generation models. Advanced controls like duration, aspect ratio, and model selection. Templates to guide new users.
Video Generator
Turn text or images into videos using the latest generation models. Advanced controls like duration, aspect ratio, and model selection. Templates to guide new users.
Batch Editor
Batch Editor
Apply AI transformations to up to 50 images at once. Choose from curated templates or build your own presets. What would take hours manually, now takes seconds.
03.3
How I worked with the team
How I worked with the team
Feasibility
Aligned before Figma opened
Design, growth, and engineering discussed the expected flow of each environment, what would and wouldn't be included, and how users would be routed there.
Work in Figma only began once all three had reached consensus.
Feasibility
Aligned before Figma opened
Design, growth, and engineering discussed the expected flow of each environment, what would and wouldn't be included, and how users would be routed there.
Work in Figma only began once all three had reached consensus.
Feedback loops
Direction checked against reality
Prototype directions were shared back with growth and engineering to confirm the design was accurately reflecting both product direction and real engineering constraints, before moving further into detail.
Feedback loops
Direction checked against reality
Prototype directions were shared back with growth and engineering to confirm the design was accurately reflecting both product direction and real engineering constraints, before moving further into detail.
Handoff
Phased, not batched
A full design handoff: screens, flows, states, and notes, was handed to engineering in phases. Once the Image Generator design was ready, it was handed off while work on the next environment continued.
This let both teams work in parallel instead of waiting on each other.
Handoff
Phased, not batched
A full design handoff: screens, flows, states, and notes, was handed to engineering in phases. Once the Image Generator design was ready, it was handed off while work on the next environment continued.
This let both teams work in parallel instead of waiting on each other.
Testing
Testing
Two lenses, one flow
Two lenses, one flow
A tester checked logic and test cases while I checked the same flow from an actual user's perspective, down to the finer details , catching different classes of issues before release.
A tester checked logic and test cases while I checked the same flow from an actual user's perspective, down to the finer details , catching different classes of issues before release.
04
Results
Results
$
$
+
+
Monthly recurring revenue
Increased by 10x
Monthly recurring revenue, increased by 10x
+
+
Monthly active users
Increased by 4x
Monthly active users, increased by 4x
%
%
Monthly revenue is now powered by
the image and video generator alone
Monthly revenue is now powered by the image and video generator alone
%
%
Increase in SEO
funnel conversions
Increase in SEO funnel conversions
End to end creative platform
Imagine, create and edit
Photorealistic output
State of the art generation models
05
More case studies coming soon
More case studies coming soon
Pricing that actually converts
Rebuilding Pixelbin's pricing funnel to increase revenue and checkout percentage
Case study coming soon
Pricing that actually converts
Rebuilding Pixelbin's pricing funnel to increase revenue and checkout percentage
Case study coming soon
Pricing that actually converts
Rebuilding Pixelbin's pricing funnel to increase revenue and checkout percentage
Case study coming soon
Pricing that actually converts
Rebuilding Pixelbin's pricing funnel to increase revenue and checkout percentage
Case study coming soon
Fashion Factory app for Reliance Jio
Integrating features like Scan and Go, daily offers and virtual aisle to enhance customer experience
Case study coming soon
Fashion Factory app for Reliance Jio
Integrating features like Scan and Go, daily offers and virtual aisle to enhance customer experience
Case study coming soon
Fashion Factory app for Reliance Jio
Integrating features like Scan and Go, daily offers and virtual aisle to enhance customer experience
Case study coming soon
Fashion Factory app for Reliance Jio
Integrating features like Scan and Go, daily offers and virtual aisle to enhance customer experience
Case study coming soon
Meet Pixie: Pixelbin's AI design agent
Building an AI design agent to help users with complex media generation usecases
Case study coming soon
Meet Pixie: Pixelbin's AI design agent
Building an AI design agent to help users with complex media generation usecases
Case study coming soon
Meet Pixie: Pixelbin's AI design agent
Building an AI design agent to help users with complex media generation usecases
Case study coming soon
Meet Pixie: Pixelbin's AI design agent
Building an AI design agent to help users with complex media generation usecases
Case study coming soon
Let's Connect
Let's Connect
Let's Connect
02
Who we designed for
Marketing leads
Needs fresh campaign creative fast, without briefing a designer or photographer each time
New to AI generation tools, so expects something closer to typing an idea than operating software
Judges the product on output quality and speed, not on feature depth
WHAT THIS MEANT FOR PIXELBIN
Navigation had to lead with "what can I make" rather than "which tool do I open," and menu labels had to stay in plain, non-technical language.
Marketing leads
Needs fresh campaign creative fast, without briefing a designer or photographer each time
New to AI generation tools, so expects something closer to typing an idea than operating software
Judges the product on output quality and speed, not on feature depth
WHAT THIS MEANT FOR PIXELBIN
Navigation had to lead with "what can I make" rather than "which tool do I open," and menu labels had to stay in plain, non-technical language.
Content managers
Handles high-volume production, needs the same edit applied across hundreds of SKUs
Came to PixelBin for specific tools like background removal, upscaling, not generation
Thinks in outputs and batches, not prompts or models
WHAT THIS MEANT FOR PIXELBIN
Batch Editor had to absorb AI features without losing the bulk-processing workflow this user already relied on
Content managers
Handles high-volume production, needs the same edit applied across hundreds of SKUs
Came to PixelBin for specific tools like background removal, upscaling, not generation
Thinks in outputs and batches, not prompts or models
WHAT THIS MEANT FOR PIXELBIN
Batch Editor had to absorb AI features without losing the bulk-processing workflow this user already relied on
Agencies & Designers
Juggles several client accounts and needs to switch context quickly
Already fluent in creative tools, wants fine-grained control over models and parameters
Evaluates PixelBin against dedicated AI generation tools, not just other editing suites
WHAT THIS MEANT FOR PIXELBIN
The generation studios needed advanced controls (model choice, duration, aspect ratio) surfaced for power users, not hidden behind simplified defaults
Agencies & Designers
Juggles several client accounts and needs to switch context quickly
Already fluent in creative tools, wants fine-grained control over models and parameters
Evaluates PixelBin against dedicated AI generation tools, not just other editing suites
WHAT THIS MEANT FOR PIXELBIN
The generation studios needed advanced controls (model choice, duration, aspect ratio) surfaced for power users, not hidden behind simplified defaults
03.3
My role
My role
Rethink
Rethink all the navigation and architecture to position AI generation as the platform’s primary capability
Redesign
Redesign the existing studio environments to cater to image and video generation and fit them in the existing console
Rebuild
Rebuild the Batch Editor, the platform’s primary power-user editing environment, to include AI features
03.3
My role
Feasibility
Aligned before Figma opened
Design, growth, and engineering discussed the expected flow of each environment, what would and wouldn't be included, and how users would be routed there.
Work in Figma only began once all three had reached consensus.
Feedback loops
Direction checked against reality
Prototype directions were shared back with growth and engineering to confirm the design was accurately reflecting both product direction and real engineering constraints, before moving further into detail.
Handoff
Phased, not batched
A full design handoff: screens, flows, states, and notes, was handed to engineering in phases. Once the Image Generator design was ready, it was handed off while work on the next environment continued.
This let both teams work in parallel instead of waiting on each other.
Testing
Two lenses, one flow
A tester checked logic and test cases while I checked the same flow from an actual user's perspective, down to the finer details , catching different classes of issues before release.