Auditing a gen-AI, interior design tool
Auditing a gen-AI, interior design tool
Auditing a gen-AI, interior design tool
UX testing project to enable the decrease of the drop-off rate during the design generation process.
UX testing project to enable the decrease of the drop-off rate during the design generation process.
Wireframe
Wireframe
Prototype
Prototype
UX Strategy
UX Strategy
ROLE
ROLE
UX Researcher
UX Researcher
UX Researcher
DOMAIN
DOMAIN
Design tool
Design tool
Design tool
TYPE
TYPE
Mixed-Method Testing
Mixed-Method Testing
Mixed-Method Testing
DURATION
DURATION
1 Month
1 Month
1 Month
Fable*, an established Interior Design SaaS startup was facing worrying levels of user drop-off.
We worked closely with them for a month to audit their product to discover the exact reason why and to recommend solutions for the same. At the end of our engagement the platform had transformed into a more functional tool that garned interest in pilot tests with target users.
Fable*, an established Interior Design SaaS startup was facing worrying levels of user drop-off.
We worked closely with them for a month to audit their product to discover the exact reason why and to recommend solutions for the same. At the end of our engagement the platform had transformed into a more functional tool that garned interest in pilot tests with target users.
Fable*, an established Interior Design SaaS startup was facing worrying levels of user drop-off.
We worked closely with them for a month to audit their product to discover the exact reason why and to recommend solutions for the same. At the end of our engagement the platform had transformed into a more functional tool that garned interest in pilot tests with target users.
User needs.
User needs.
User needs.
Fig 1: Home Owners (For representation purposes only)
Fig 1: Home Owners (For representation purposes only)
Fig 1: Home Owners (For representation purposes only)
The web application promised to bring home design visions to life for amateur creatives and homeowners alike.
With the help of Fable's cutting edge Generative AI Engine, users can generate images and make spot-edits using their magic edit features.
The web application promised to bring home design visions to life for amateur creatives and homeowners alike.
With the help of Fable's cutting edge Generative AI Engine, users can generate images and make spot-edits using their magic edit features.
The web application promised to bring home design visions to life for amateur creatives and homeowners alike.
With the help of Fable's cutting edge Generative AI Engine, users can generate images and make spot-edits using their magic edit features.
Key Challenges.
Key Challenges.
Key Challenges.
Despite a clear business model and interesting features, a few critical issues emerged:
Despite a clear business model and interesting features, a few critical issues emerged:
Despite a clear business model and interesting features, a few critical issues emerged:
Users were dropping off without completing the full flow.
Users were dropping off without completing the full flow.
Users were dropping off without completing the full flow.
The platform overwhelmed the users with excessive tool choices.
The platform overwhelmed the users with excessive tool choices.
The platform overwhelmed the users with excessive tool choices.
The image generation process was linear and time-consuming.
The image generation process was linear and time-consuming.
The image generation process was linear and time-consuming.
The Discovery Process.
The Discovery Process.
The Discovery Process.
Pilot interviews.
Pilot interviews.
Pilot interviews.
First, we needed to understand how the platform is viewed by the target audience. Early interviews revealed that while many would consider such a tool as a great starting point for ideation, it cannot fully substitute professional design intervention / tools. This was especially true for creatives.
This early concept validation helped us stabilize our trajectory in two ways:
Design of in-depth testing - the questions we designed to probe our audiences shifted slightly knowing where the platform fit into their design process
Solution space - During later stages, this line of thought also influenced how we demoed the product to focus groups and sought ideas.
First, we needed to understand how the platform is viewed by the target audience. Early interviews revealed that while many would consider such a tool as a great starting point for ideation, it cannot fully substitute professional design intervention / tools. This was especially true for creatives.
This early concept validation helped us stabilize our trajectory in two ways:
Design of in-depth testing - the questions we designed to probe our audiences shifted slightly knowing where the platform fit into their design process
Solution space - During later stages, this line of thought also influenced how we demoed the product to focus groups and sought ideas.
First, we needed to understand how the platform is viewed by the target audience. Early interviews revealed that while many would consider such a tool as a great starting point for ideation, it cannot fully substitute professional design intervention / tools. This was especially true for creatives.
This early concept validation helped us stabilize our trajectory in two ways:
Design of in-depth testing - the questions we designed to probe our audiences shifted slightly knowing where the platform fit into their design process
Solution space - During later stages, this line of thought also influenced how we demoed the product to focus groups and sought ideas.
Market position.
Market position.
Market position.
A competitor analysis helped us identify gaps in Fable's offerings:
Editing objects - Manipulation of furniture was not enabled
Collage options - Users did not have the option of creating moodboards
A competitor analysis helped us identify gaps in Fable's offerings:
Editing objects - Manipulation of furniture was not enabled
Collage options - Users did not have the option of creating moodboards
A competitor analysis helped us identify gaps in Fable's offerings:
Editing objects - Manipulation of furniture was not enabled
Collage options - Users did not have the option of creating moodboards
Fig 2: Competitor Analysis
Fig 2: Competitor Analysis
Fig 2: Competitor Analysis
Fig 3: Heuristic analysis
Fig 3: Heuristic analysis
Fig 3: Heuristic analysis
Heuristic evaluation.
Heuristic evaluation.
Heuristic evaluation.
An in-depth walkthrough and evaluation were key in understanding the high-drop off rates:
Poor navigational flow - The flow towards a design output was rigid and complex
Uncanny valley - The quality of the generated imaged were slightly unrealistic & off-putting
An in-depth walkthrough and evaluation were key in understanding the high-drop off rates:
Poor navigational flow - The flow towards a design output was rigid and complex
Uncanny valley - The quality of the generated imaged were slightly unrealistic & off-putting
An in-depth walkthrough and evaluation were key in understanding the high-drop off rates:
Poor navigational flow - The flow towards a design output was rigid and complex
Uncanny valley - The quality of the generated imaged were slightly unrealistic & off-putting
Moderated user testing.
Moderated user testing.
Moderated user testing.
We conducting usability test on target users. They led to two surprising discoveries that we hadn't accounted for:
Credit anxiety problem - A psychological pattern emerged during testing. Participants repeatedly expressed discomfort about the visible depletion of the credits from even trivial actions like deleting furniture. Based on this insight, we sought to revise the credit model.
AI "black box" problem - Participants also exressed confusion about the Gen AI process. The problem we realized was due to the confusing signals for it's various states , and the lack of clear documentation.
We conducting usability test on target users. They led to two surprising discoveries that we hadn't accounted for:
Credit anxiety problem - A psychological pattern emerged during testing. Participants repeatedly expressed discomfort about the visible depletion of the credits from even trivial actions like deleting furniture. Based on this insight, we sought to revise the credit model.
AI "black box" problem - Participants also exressed confusion about the Gen AI process. The problem we realized was due to the confusing signals for it's various states , and the lack of clear documentation.
We conducting usability test on target users. They led to two surprising discoveries that we hadn't accounted for:
Credit anxiety problem - A psychological pattern emerged during testing. Participants repeatedly expressed discomfort about the visible depletion of the credits from even trivial actions like deleting furniture. Based on this insight, we sought to revise the credit model.
AI "black box" problem - Participants also exressed confusion about the Gen AI process. The problem we realized was due to the confusing signals for it's various states , and the lack of clear documentation.
Solution Space.
Solution Space.
Solution Space.
Workshopping session.
Workshopping session.
Workshopping session.
We took all the problems we had identified to a panel of Senior UX Designers for a brainstorming session. This focused discussion helped us generate breakthrough feature ideas:
Decrease friction - One significant breakthrough was taking inspiration from Canva - keeping all design tools usable in one canvas-like area.
We took all the problems we had identified to a panel of Senior UX Designers for a brainstorming session. This focused discussion helped us generate breakthrough feature ideas:
Decrease friction - One significant breakthrough was taking inspiration from Canva - keeping all design tools usable in one canvas-like area.
We took all the problems we had identified to a panel of Senior UX Designers for a brainstorming session. This focused discussion helped us generate breakthrough feature ideas:
Decrease friction - One significant breakthrough was taking inspiration from Canva - keeping all design tools usable in one canvas-like area.
Stakeholder Presentation .
Stakeholder Presentation.
Stakeholder Presentation.
We finally presented six key areas of improvement to Fable's leadership, each with clear rationale and action points.
We finally presented six key areas of improvement to Fable's leadership, each with clear rationale and action points.
We finally presented six key areas of improvement to Fable's leadership, each with clear rationale and action points.
Fig 4: Stakeholder Receommendations
Fig 4: Stakeholder Receommendations
Fig 4: Stakeholder Receommendations
We used a 'Cost vs Benefit' matrix to highlight the recommendations that would bring the highest value, to help Fable prioritize effectively while keeping the trade-off manageable.
While we recognized potential constraints, especially technical ones in the Gen AI space for the time, we flagged the issues for future resolution.
We used a 'Cost vs Benefit' matrix to highlight the recommendations that would bring the highest value, to help Fable prioritize effectively while keeping the trade-off manageable.
While we recognized potential constraints, especially technical ones in the Gen AI space for the time, we flagged the issues for future resolution.
We used a 'Cost vs Benefit' matrix to highlight the recommendations that would bring the highest value, to help Fable prioritize effectively while keeping the trade-off manageable.
While we recognized potential constraints, especially technical ones in the Gen AI space for the time, we flagged the issues for future resolution.
Post-implementation.
Post-implementation.
Post-implementation.
Fig 5: Before & After our intervention
Fig 5: Before & After our intervention
Fig 5: Before & After our intervention
The Fable team put our recommedations to work based on our comprehensive testing report.
We also provided a second round of evaluation, this time ensuring a system wide consistency in design language, UX copy and mental models.
The result was a clean and clear interface that pilot audiences operated without showing any signs of drop-off (as per the anecdotes shared with us by Fable). At the end of our engagement, Fable had a fully transformed creative SaaS tool that supported their vision.
The Fable team put our recommedations to work based on our comprehensive testing report.
We also provided a second round of evaluation, this time ensuring a system wide consistency in design language, UX copy and mental models.
The result was a clean and clear interface that pilot audiences operated without showing any signs of drop-off (as per the anecdotes shared with us by Fable). At the end of our engagement, Fable had a fully transformed creative SaaS tool that supported their vision.
The Fable team put our recommedations to work based on our comprehensive testing report.
We also provided a second round of evaluation, this time ensuring a system wide consistency in design language, UX copy and mental models.
The result was a clean and clear interface that pilot audiences operated without showing any signs of drop-off (as per the anecdotes shared with us by Fable). At the end of our engagement, Fable had a fully transformed creative SaaS tool that supported their vision.
Key Takeaways.
Key Takeaways.
Key Takeaways.
This project taught me the value of staying light on my feet, and improvising as the work changes trajectories. This came about when we realized that our original research plan was no longer serving the needs of the project and we shifted gears accordingly.
This project taught me the value of staying light on my feet, and improvising as the work changes trajectories. This came about when we realized that our original research plan was no longer serving the needs of the project and we shifted gears accordingly.
This project taught me the value of staying light on my feet, and improvising as the work changes trajectories. This came about when we realized that our original research plan was no longer serving the needs of the project and we shifted gears accordingly.
Get in touch with me for opportunities, consultations or even just a chat
ALL RIGHTS RESERVED. ANJANA ELANGOVAN 2025
Get in touch with me for opportunities, consultations or even just a chat
ALL RIGHTS RESERVED. ANJANA ELANGOVAN 2025
Get in touch with me for opportunities, consultations or even just a chat
ALL RIGHTS RESERVED. ANJANA ELANGOVAN 2025