Colour with Asian Paints

Increase Lead Generation via
Augmented Reality based
AI Designer tool

Android & iOS

Augmented Reality Gen AI Nano Banana Segment Anything Model (SAM) by META Home AI Design

Summary Snapshot

Before → After

Before Business Problems:

Downward trend of lead generation in Colour with Asian Paints app’s Visualiser.
High drop off noticed on Adobe Analytics during selection of Paints/Textures/Wallpapers in the visualisation journey
Low CTR on lead generation form-fill CTA.

User Pain-points:

5000+ options in our product catalogue caused decision paralysis.
Traditional “browse & choose” interaction formed a barrier between users’ imagination & the catalogue, leading to unsatisfactory visualiser result.
Lack of value communicated through form-fill CTA label made the user feel uncertain.

After Design Decisions:

Introduced 2 highlighted entry points to increase feature CTR, along with simple flow to make AI adoption quick & intuitive.
Cleary distinguished input methods: text & via image based, informed the users of all tech literacy backgrounds about available ways to interact with AI.
Result page shows which products were used. Manual override present to always keep user at final control.
Paint budget calculation as lead gen activity provided value based motive and certainty to impact CTR.

The Product

Its a digital visualisation app allowing over 5 million users to instantly view, select & visualise over 5000+ paint shades, textures & wallpapers on their own wall before painting. It also acts as a lead generation and engagement app for various products & services of Asian Paints.

The Users

The target group is of 5M + users ranging from 25 to 45+ years of age concentrated mostly in tier 1 & 2 Indian cities typically having mid to high tech literacy. It is focussed mainly towards those who are planning to repaint their homes.

0M+

users

view, select & visualise

paint shades, textures & wallpapers

on their own wall before painting

25 – 45+

years of age

tier 1 & 2 Indian cities

My Role

Lead Product Designer.

Owned the journey end-to-end.

Requirements Research Product and design strategy Wireframes UI Validation Dev handover Hand off docs

XX weeks explorations YY weeks design alignment

I partnered with the team throughout the process, we had regular check-ins, to iterate quickly and ensure we were building something usable and aligned with the product vision. It was a fast moving team & we stayed tightly aligned from research & exploration through hand off.

Context

The visualiser feature inside the Colour with Asian Paints (CWAP) app is a home visualisation tool that allows users to visualise how any paint/wallpaper/texture shades/styles of their choice would look like on their walls after application. This is also a lead generation tool to pitch home painting services offered by Asian Paints.

Lately lead generation showed a downward trend with increased drop off rate. My goal as a UI/UX designer was to redesign the tool to boost lead generation. (This is an ongoing project)

Problem

The challenge was to increase user satisfaction in the whole paint selection decision making process. Paint selection can be cumbersome.

If the journey to selecting paints/textures/wallpapers is smooth, it would automatically reduce drop off rates and increase lead generation.

User Pain Points

One category of people decide a particular shade/style and then visit the app to visualise it ion their wall.

The other category of people visits the app without any particular shade/style of paint/texture in mind.

In either cases, arriving at a particular shade/style they like is a tedious process as colour catalogues offer vast choices & very often lead to decision paralysis resulting in drop offs.

People painting their home usually have many ideas and references in mind but struggle to execute them by selecting shades from catalogues. Also, visions in imagination real world outcomes. This causes disappointment.

Customers after deciding upon a shade would typically like to know how much would the painting cost be. But that journey was separate & quite long.

The Solution

To help users choose paint shades that match their imagination and not get exhausted in the vast sea of 5000+ shades, AI in the visualiser would be the perfect tool!

The model is being trained on the complete catalogue of Asian Paint’s 5000+ paint shades, texture & wallpaper styles.

Provide a way for the customer to calculate the painting budget. This meaningful insight would further convince the customer to opt for painting services by providing relevant information to make an informed choice. Thus providing linking to the Paint Budget Calculator journey from the AI Visualiser result page has a high potential to boost leads generated.

KPIs with targets

−0%

Completion Rate

Decrease mid journey drop offs by 50%

+0%

Lead Generation

Increase lead generation by 45%

+0%

Downloads

Increase app downloads by 30%

Task Flow

Ideation

The design philosophy revolved around effortless simplicity & faster decision making. Every interaction was crafted to minimise friction and guide users naturally through the paint shade selection process.

Once the user was satisfied with their result, lead generation CTAs could also be introduced strategically.

Entry points

Initial Design Directions

In the early stages, I explored different layout directions, to find the right balance between clarity and emphasis. The idea was to provide a prominent message to the users and design convenient entry points for them to access the AI visualiser. I decided to design a hero banner and place it at the top, along with the existing entry point from the bottom navigation.

Get started

Simple ways to get started
in the visualisation journey.

The journey began with a selection of space type: Interior/Exterior, followed by uploading an image of the desired space. All the options were put up front and prominently which made the journey simple and familiar.
Space type: Interior/Exterior values were decided to be gathered at the beginning in order to reduce the clutter and complexity of the image selection/upload page in the next step and also so it would reduce chances of error by the model in case, it fails to detect the space type.
Templates were included to help the users get started if they were unable to upload or shoot an image of their own space due to any circumstances. This would potentially prevent drop offs.

A change button was included in the preview/prompt input page. This provided an easy out to the users to recover from the mistake of choosing the wrong image or incase they changed their minds.

Describe vision

Effortless and flexible methods
to describe painting ideas/vision.

The preview/input page contained two interaction points which allows the users to provide directions to the AI model. This was based on a few observations:

It is easier to describe an idea than looking for items that match this idea from a catalogue of 5000+ shades.

It is easier to describe this idea through visual references.

Expecting every user to realise the usefulness of a visual reference is a far fetch.

Thus, providing the option to upload a reference image in the same input field as the prompt, would not have the desired impact as it can be easily overlooked by users with lower tech literacy.

Hence, having a dedicated button upfront allowing them to upload a reference image was a more inclusive & universal design approach.

Input via Prompt

Input via Reference Image

Processing

Adapting to the limitations
of the technology.

Processing the image with AI required time. It could take approximately little over a minute to generate the results. This could lead to potential drop offs in an otherwise simple and quick journey. To counter this issue, users were presented with two options:

Stay and wait for the result to appear.

Go to home and explore other features, get informed when result is ready.

1. Stay and Wait

2. Go home and be notified

If the user chooses to go to home page and wait, exploring other features of the app, the system keeps them informed of the progress. The visualise button in the bottom nav was adapted to show this message and designed such that it did not over shadow or intrude into other tasks and only serve its purpose, quietly.
We also handled a scenario if the user grew impatient while the result is being generated by providing a pop-up conveying the message that result is being worked upon in case they tap on the visualise button while it is processing.
If they went to the device’s home screen or another app, they would be notified once the result is ready via device notification.

Result

Simple result page with multiple
options and relevant features.

A straight & simple result page was designed which showed the result image clearly with prominence and all the other information centring this result image making the design intuitive and having correct information hierarchy.

Result options were put upfront with a button placed right next to them allowing the user to generate more options.
Controls such as: Save, Download, Before-After and Restart journey button group was placed right below.
Shades used were clearly mentioned with codes for easy reference to be taken when/if they decide to purchase.
CTA to cost calculation was placed strategically next to the shades mentioned, nudging users to satisfy their budgetary curiosity, thus completing the requirement of a paint buyer’s logical needs in a single loop.
A prominent lead generation button was placed below for high visibility.

Customise

Flexibility to customise
the generated results.

Communication is key! Result will be as good as the prompt communicated to the AI model. This Prompt is the Key

It is quite understandable that users might not always be able to express themselves completely and thus, there could arise some disparity between their expectations and real world generated result.
To further increase usability, a manual over ride was necessary. this makes the user feel & stay in control of the system.
The manual change of colours allows the users to find and select nearest alternate shades. If that doesn’t help, they can go looking at the complete colour catalogue.

Current challenge

Since the model is unable to identify which colour has been used on which wall, changing a colour is not generating the expected result.

Closing notes.

01Baseline Sentiment+

Lead generation in the visualiser was declining because users were dropping off during paint, texture, and wallpaper selection. From the user’s side, the journey felt overwhelming, slow, and difficult to translate into a confident decision.

02AI-guided discovery to solve choice overload+

The 5000+ catalogue created decision paralysis, so the experience shifts from browsing-heavy selection to an AI-assisted visualiser, supported by a simple start flow, guiding nudges, and templates.

03Flexible inputs to smoothen idea-to-shade translation+

People often think in moods, ideas, or references rather than exact shade names; multiple input methods were introduced such as prompts and reference-based inputs for different comfort levels.

04Increased user autonomy to boost confidence+

In an event of unsatisfactory result, trust drops. More control was given to users through a simplified results page with clear output, manual override controls, and a shorter visualise-to-budget loop.

05Value perception to overcome uncertainty+

After users overlooked the lead generation form-fill CTA, value was introduced by “Calculate Budget” feature embedding a form-fill to act as a nudge to increase form-fill CTR.

06Impact+

By making the journey shorter, clearer, and more supportive, the redesigned AI visualiser is expected to reduce drop-offs, improve confidence in the output and the brand, and strengthen lead-generation potential.