Introducing Deco: AI interior designer

Introducing Deco: AI interior designer

Deco on iOS is an AI interior design app that allows users to upload photos of their interior and redesign it using generative AI tools. It is live on the app store and available in the US and Canada.


My work: I was the sole designer working on this project. I worked on everything from user research, to conceptualization ,wireframing and prototyping, all the way up until user testing and releasing the app on iOS

Deco on iOS is an AI interior design app that allows users to upload photos of their interior and redesign it using generative AI tools. It is live on the app store and available in the US and Canada.


My work: I was the sole designer working on this project. I worked on everything from user research, to conceptualization ,wireframing and prototyping, all the way up until user testing and releasing the app on iOS

My role

My role

Product Designer

Product Designer

Project

Project

Professional work

Professional work

Timeline

Timeline

4 months (Jan - July 2024)

4 months (Jan - July 2024)

Tools used

Tools used

Figma, LottieFiles Adobe Illustrator, SVGator

Figma, LottieFiles Adobe Illustrator, SVGator

"Hmm… I wonder how much it would cost for an Interior Designer to redesign my living room…."

"Hmm… I wonder how much it would cost for an Interior Designer to redesign my living room…."

"NO WAY!? It's way too expensive! There's got to be a way I can do this on my own…."

"NO WAY!? It's way too expensive! There's got to be a way I can do this on my own…."

What was the problem?

What was the problem?

What was the problem?

  • Interior designers charge between $100-$1500 for simple conceptual moodboards and renders of interior spaces


  • There weren't any easy to use DIY interior design tool on the market

  • Interior designers charge between $100-$1500 for simple conceptual moodboards and renders of interior spaces


  • There weren't any easy to use DIY interior design tool on the market

We surveyed a 100 homeowners across the US and Canada

We surveyed a 100 homeowners across the US and Canada

We surveyed a 100 homeowners across the US and Canada

60

Said they were willing to design their interiors themselves

Said they were willing to design their interiors themselves

76

Had taken on a small DIY interior project at home

Had taken on a small DIY interior project at home

What does this mean?

What does this mean?

  • A lot of people are willing to design interiors themselves, they just need better tools.


  • By delivering intuitive, well-designed tools, we empower users to independently achieve professional designer-like results.

  • A lot of people are willing to design interiors themselves, they just need better tools.


  • By delivering intuitive, well-designed tools, we empower users to independently achieve professional designer-like results.

Research summary

Research summary

Secondary research

Fill in the details for calculation

Literature reviews

Literature reviews

Competitive analysis

Competitive analysis

Market trends analysis

Market trends analysis

Cognitive task analysis

Cognitive task analysis

Primary research

Fill in the details for calculation

User interviews

User interviews

Usability testing (2 rounds)

Usability testing (2 rounds)

Affinity Mapping

Affinity Mapping

Prototyping & testing

Fill in the details for calculation

Wirefraaming

Wirefraaming

Design system

Design system

High-Fidelity prototyping

High-Fidelity prototyping

Usability testing

Usability testing

Competitive analysis

Competitive analysis

Competitive analysis

Analysis of direct and indirect competitor apps revealed key gaps and unmet user needs:

Analysis of direct and indirect competitor apps revealed key gaps and unmet user needs:

Too many features, barely any functional

Too many features, barely any functional

Poor UX, glitchy interface

Poor UX, glitchy interface

More modelling and preset features that required a learning curve

More modelling and preset features that required a learning curve

Competitor app reviews

Competitor app reviews

  • A lot of good features were behind a paywall with high subscription costs

  • Very glitchy interfaces

  • Lots of negetive reviews indicating dissatisfaction

  • A lot of good features were behind a paywall with high subscription costs

  • Very glitchy interfaces

  • Lots of negetive reviews indicating dissatisfaction

Potential solutions:

Potential solutions:

Generative AI




Fill in the details for calculation

With the recent advancements in Generative AI, users can explore designs and create what they imagine

With the recent advancements in Generative AI, users can explore designs and create what they imagine

Small learning curve



Find the mortgage calculator

Focus on features that don’t need a learning curve targeted to amateur designers

Focus on features that don’t need a learning curve targeted to amateur designers

Strong UX




Strong UX

Need to make sure that app is very easy to use and is amateur designer centric.

Need to make sure that app is very easy to use and is amateur designer centric.

The market gap and strategic direction became clear, guiding the next steps for targeted design solutions…

The market gap and strategic direction became clear, guiding the next steps for targeted design solutions…

The market gap and strategic direction became clear, guiding the next steps for targeted design solutions…

Wireframing

Wireframing

Wireframing

  1. First, I tried to draw out 3d environment where users can drag and drop furniture and move it around on different axes to place it however they liked.

    We did not end up going with this approach due to:

    • Learning curve

    • Time and technical constraints

  1. First, I tried to draw out 3d environment where users can drag and drop furniture and move it around on different axes to place it however they liked.

    We did not end up going with this approach due to:

    • Learning curve

    • Time and technical constraints

  1. Second, I sketched out a flow where users could upload a picture and use and Generative AI chat interface to make modifications like - "Add a painting on the wall."

    This was a good approach:

    • Easy to learn

    • First time mover advantage


This is the flow used that was finalized for the MVP

  1. Second, I sketched out a flow where users could upload a picture and use and Generative AI chat interface to make modifications like - "Add a painting on the wall."

    This was a good approach:

    • Easy to learn

    • First time mover advantage


This is the flow used that was finalized for the MVP

  1. I felt the need for users to draw inspiration from somewhere and take it a step further by letting them edit it exactly to their preferences and tastes.


    So a user could browse images > Select an image > Modify it using GenAI > Save and share it


In the interest of time, this feature was deprioritized and not included for the MVP launch.

The MVP prioritized intuitive GenAI features to ensure ease of use and maximize user value.

The MVP prioritized intuitive GenAI features to ensure ease of use and maximize user value.

The work on the MVP commenced…

The work on the MVP commenced…

The work on the MVP commenced…

FINAL DESIGN

FINAL DESIGN

FINAL DESIGN

Just Click or Upload a photo

Just Click or Upload a photo

Just Click or Upload a photo

Click a photo of an interior space and choose the room type it belongs to.


It's that simple!

Click a photo of an interior space and choose the room type it belongs to.


It's that simple!

Design Decision: Loading animations

Design Decision: Loading animations

Design Decision: Loading animations

Upto 10 seconds

Upto 10 seconds

of loading time

of loading time

The engineering team stated that it would take upto 10 seconds to process and generate different ai interior styles.


Decision: Designed engaging loading animation screens to keep the user occupied and mask loading time.

The engineering team stated that it would take upto 10 seconds to process and generate different ai interior styles.


Decision: Designed engaging loading animation screens to keep the user occupied and mask loading time.

Type what you want. AI does the rest!

Type what you want. AI does the rest!

Type what you want. AI does the rest!

Want to add a painting on your wall? Just type and see magic happen.


Users can also select from choices generated and replace the existing design with just a tap.

Want to add a painting on your wall? Just type and see magic happen.


Users can also select from choices generated and replace the existing design with just a tap.

User Testing - Round 1

User Testing - Round 1

User Testing - Round 1

Usability Testing insights

Usability Testing insights

  • Users could not grasp all the features available


  • Most users said they find all the features useful


  • Most users said they would download and use the app

  • Users could not grasp all the features available


  • Most users said they find all the features useful


  • Most users said they would download and use the app

12

12

12

Number of users

Number of users

Number of users

70%

70%

70%

Task completion rate

Task completion rate

Task completion rate

1 min 35 secs

1 min 35 secs

1 min 35 secs

Average time per task

Average time per task

Average time per task

72.2

72.2

72.2

Average SUS

Average SUS

Average SUS

Results

Results

Results

7/12

7/12

Did not understand how to use the features

Did not understand how to use the features

10/12

10/12

Participants wanted onboarding tutorials.

Participants wanted onboarding tutorials.

Decisions to be made

  1. Design progressive onboarding that teaches new users about the different tools and how to use them


  1. Design a shop feature for users to find products that they see in their generated design

  1. Design progressive onboarding that teaches new users about the different tools and how to use them


  1. Design a shop feature for users to find products that they see in their generated design

  1. Design progressive onboarding that teaches new users about the different tools and how to use them


  1. Design a shop feature for users to find products that they see in their generated design

Design change - Progressive onboarding

Design change - Progressive onboarding

Design change - Progressive onboarding

Using a combination of flash cards and animations during onboarding, the app effectively explains how to make the most of the new tools.

Using a combination of flash cards and animations during onboarding, the app effectively explains how to make the most of the new tools.

Added feature - Shop products in your design

Added feature - Shop products in your design

Added feature - Shop products in your design

After users create their dream design, the app automatically tags furniture in the image and shows you links to similar products across the web.

A lot of these are affiliate links, so if the user makes a purchase, Deco gets a share too!

After users create their dream design, the app automatically tags furniture in the image and shows you links to similar products across the web.

A lot of these are affiliate links, so if the user makes a purchase, Deco gets a share too!

User Testing - Round 2

User Testing - Round 2

User Testing - Round 2

What was tested?

What was tested?

  • The new progressive onboarding flow


  • The new shop feature

  • The new progressive onboarding flow


  • The new shop feature

Results

Results

Results

4/5

4/5

Found the onboarding experience useful

Found the onboarding experience useful

3/5

3/5

Said they would use the shop feature

Said they would use the shop feature

Results after debuting the MVP on the App store

Results after debuting the MVP on the App store

Results after debuting the MVP on the App store

100+ Signups

100+ Signups

In the first one month of launch on the app store

In the first one month of launch on the app store

Over 60%

Over 60%

Engagement with 2 more more features

Engagement with 2 more more features

  • Deco launched on the app store in June 2024 with a limited set of shipped features. Through our aggressive social media marketing, we were able to get a considerable number of downloads.


  • In July 2024, I left the company to pursue my graduate studies.

  • Deco launched on the app store in June 2024 with a limited set of shipped features. Through our aggressive social media marketing, we were able to get a considerable number of downloads.


  • In July 2024, I left the company to pursue my graduate studies.

Key takeaways

Key takeaways

I learnt so much over the summer working with an amazing set of Designers, Developers and PMs!

I learnt so much over the summer working with an amazing set of Designers, Developers and PMs!

Designing from 0 to 1

Designing from 0 to 1

This was the first time I got the chance to be involved in a project right from the Idea stage to shipping an MVP. I got to experience the full product life cycle and improvise whenever necessary.

This was the first time I got the chance to be involved in a project right from the Idea stage to shipping an MVP. I got to experience the full product life cycle and improvise whenever necessary.

Designing around tech constraints

Designing around tech constraints

There were a lot of times tech constraints posed an obstacle to ship some design features/ elements. I had to come up with create ways to circumvent this and still deliver value to the user.

There were a lot of times tech constraints posed an obstacle to ship some design features/ elements. I had to come up with create ways to circumvent this and still deliver value to the user.

NIHAL PALOCAREN

NIHAL PALOCAREN