DreamTeam
AI-powered fantasy cricket team builder for Dream11
My Role
Product Designer
Timeline
Jan-July 2025
Scope
UX/UI design for the entire customer booking platform, including the customer-facing app, internal dashboard, and the product's scalable design system
Project Overview
What is Cleannn?
Finding trusted home cleaning in a city as fast-paced as Bangkok is a constant challenge.
Cleannn is an on-demand platform connecting users with reliable, vetted cleaning professionals.
I built the entire customer experience from the ground up—leading everything from initial research and branding to the final UX and UI of the app.
Chapter 1
The Messy Challenge
To build a digital concierge that makes booking a trusted cleaner feel effortless.
Life in Bangkok is fast, and cleaning is often the first thing people skip. With the language barrier, even booking a simple home cleaning can feel confusing and difficult.
Chapter 2
Kickoff
I began by aligning with the founder's vision to form initial hypotheses on user pain points, cultural impacts, and where design could provide the most value.
Initial Insights
Trust is the real service
Clean apartments were expected; what people truly craved was reliability and someone they could count on.
Booking shouldn't be a chore
Existing options made scheduling a cleaner feel like a chore in itself.
Language barrier
Most of the Expats faced language barrier while booking even normal home cleaning services
Chapter 3
The Research
Phase
Combining stakeholder interviews, real-world competitor analysis, and live user feedback to uncover the core problems.
Alright, so with my initial hypothesis set, it was time to validate them. So I started digging in from every angle to find patterns that would help me understand actual challanges
Competitor Analysis
I didn't just look at screenshots. We booked, rescheduled, and cancelled cleanings with our main competitors. This gave us a firsthand account of the friction and frustrations users faced across the industry.
Major Insights
Competitors hide the price
We saw that almost every competitor shows the final price only at the end of the booking. This feels like a trick and frustrates users.
Add-ons are hard to find
Important add-ons like laundry or ironing were often buried deep in the settings, making it a pain for users to customize their order.
Confusing booking flows
Many apps had confusing layouts and too many steps. We saw this as a clear chance to win by being the simplest and fastest.
User Reseach
Before the app, we had our service live on Instagram. I used it to chat with our users, run various campaigns/polls etc, to understand their key pain points, and their mental models.
Major Insights
Users want the price first
We saw that almost every competitor shows the final price only at the end of the booking. This feels like a trick and frustrates users.
Scheduling needs to be flexible
We saw a lot of last-minute reschedules. This told us we needed to offer flexible time slots and have a very clear confirmation step.
They want the same maid back
We learned that people cared way more about getting the same helper they trusted than they did about getting a small discount.
"What If I'm Not Home?"
This anxiety about logistics (like key pickups or building access) was a major point of friction that most users faced.
Chapter 4
Connecting the Dots
After synthesizing all the research, four core problems became crystal clear. These were the roadblocks we had to remove.
One-Size-Fits-All
How might we design a flexible experience that offers simple guidance for beginners while providing the deep data and control that experts demand?
Lack of Trust
How might we build user trust by designing an interface that makes AI recommendations simple, transparent, and easy to understand?
Scattered Data
How might we create a single, efficient interface that integrates all the essential stats and data users need to make a confident decision?
High-Risk Strat.
How might we provide tools for experts to build unique lineups and a practice arena to test their high-risk strategies safely?
Chapter 5
From Blueprint to Pixels
I built the product's foundation with detailed user flows, IA, and wireframes, while establishing a design system for consistency and scale.
With the core problems defined, it was time to build the solution. My process was highly iterative and collaborative, focusing on translating insights into a tangible, user-friendly experience.
User Flow and IA
Using insights from Instagram user research, chats, and competitor analysis, I mapped a simple, linear user flow reflecting real user behavior. Through card sorting, I structured a scalable information architecture that minimized friction from the very first screen.
What's Changed in the new user flow?
Introducing "Price Calculator"
To solve for one of the biggest pain point, "hidden prices", we added a "Price Calculator" right at the starting, so that users can see if the service is economically feasible for them, this generates trust and grabs users attention, and hence reducing drop offs.
Add-ons are easy to find
Important add-ons like laundry or ironing were made visible and users can directly add them in a single click right in the first step. It became a part of the price calculator.
Streamlined and Easy booking flow
Users can book a service in just 4 easy steps. No hidden charges, no iterative flow.
Wireframing
I sketched out different ways in which we can solve our problems. Creating different flows and approaches. I chose to use pen and paper for wireframing as it helped me quickly ideate many different Ideas and because I had my design system ready, I made quick UI to discuss with the team and validate my designs.
Anticlimax
The Design System
Pivot
To move faster, we pivoted from a custom design system to modifying an existing one, saving dev time while keeping our brand consistent.
To ensure consistency, I initially built a custom design system from the ground up. However, we quickly realised that maintaining a fully custom system would slow down development. So, we made a pragmatic pivot. We adopted and heavily modified the "Reshaped Design System" to match our unique branding.
Chapter 6
The Solution
A user-friendly platform that solves for complexity with a simple, trustworthy design.
With our validated blueprint and our design system in place, I moved forward to craft the final solution.
Booking Flow
Booking a cleaner shouldn't feel like chore.
We replaced the confusing maze with a simple, linear path. Users just select their service, choose a time, and confirm
Video Preview
Video Preview
Subscription flow
No more manual rebookings everytime
Users can now schedule their favorite service to repeat weekly or monthly, securing their preferred time slot and getting a clean home on autopilot—no more manually booking chores.
Transparent Pricing
We decided to show prices upfront to build trust with users
Users now see a clear, upfront cost based on their service and area before they even enter the booking flow. This builds immediate trust, eliminates the fear of hidden fees, and lets the user decide to proceed with full confidence.
Video Preview
Video Preview
Rebooking made easy
Rebook last service directly from homepage
For returning customers, a "Book Again" button appears on homepage, letting them rebook their most recent service in a single tap—no forms, no fuss, just instant confirmation.
Chapter 7
The Impact
฿320,000
Total Revenue Generated within first 8 months of starting
522
Total bookings
฿101,049
Peak monthly revenue (July 2025)
Final Chapter
Retrospective
Working on Cleannn from scratch was one of the most defining experiences in my design journey.
As the only designer in a small team of three developers, I helped turn an idea into a working product. With limited time and resources, we focused on launching an MVP fast and kept improving it based on real user feedback.
Key Learnings
Learning to Be Practical
I initially started to building a full custom design system, but speed mattered more than perfection. With a small team and tight deadlines, I customised an existing system, "Reshaped"— a valuable lesson in balancing ambition with reality.
Owning the Whole Process
From user chats to final UI, I handled the entire design process. It made me more resourceful, confident, and aware of how every detail shapes the user experience.
Seeing Real Impact
Watching bookings and revenue rise after each month was the best validation — proof that good design, even with constraints, can drive real business results.
Blurring the Line Between Design and Business
Here I learnt that great design goes beyond visuals — it’s about understanding operations, users, and growth. Working closely with founders and teams, I learned how design decisions can directly shape strategy, trust, and scalability.
DreamTeam
AI-powered fantasy cricket team builder for Dream11
My Role
Product Design lead
Timeline
Nov 2024
Scope
Research, design, and prototype the complete user experience for the DreamTeam to the final UI handoff for development.
Project Overview
By merging a predictive ML model with a user-centric design, our solution simplified strategic team creation and won a bronze medal.
For the Inter-IIT Tech Meet, our team developed DreamTeam, an AI-powered fantasy cricket team builder for Dream11.
As Design Lead, I guided a team of three designers through the entire product design lifecycle, from foundational research and strategy to final handoff for development.
The Problem
Fans struggle to quickly build trustworthy, data-driven fantasy teams due to scattered information and low confidence in AI recommendations.
Our research showed that users struggle to find a reliable, all-in-one platform for quick team creation. They feel overwhelmed by scattered data and skeptical about AI-driven recommendations, fearing unclear logic behind suggestions that impact their performance and money.
Is it really worth solving?
220 Million
Dream11 users, the world's largest fantasy sports platform
31%
The massive growth rate of the fantasy sports industry in India during FY23
₹2.19 Lakh Crore
The projected market size of the Indian fantasy sports industry by FY27
The Solution
Introducing DreamTeam, your personal AI strategist designed to make the complex art of fantasy team creation simple, transparent, and quick.
DreamTeam builds trust by explaining the "why" behind every AI pick and allows users to customize the level of AI assistance they desire, ensuring they are always in control.
Chapter 1
How It All Began
It all started with an email. The problem statements for the Inter-IIT Tech Meet 13.0, released by IIT Bombay, were out. Among them was a compelling challenge from Dream11
A multidisciplinary product team was formed at IIT Guwahati, and my first step as the incoming Design Lead was to recruit and lead a dedicated team of three designers to tackle the user experience.
Chapter 2
Understanding the Landscape
Our strategy was to first understand the complete ecosystem—the competition's goals, the product's current state, and the market's key players—before we could ask users the right questions.
With our design team in place, our first step was to build a 360-degree view of the challenge. We began with a deconstructing the competition brief, performing a complete product walkthrough of the Dream11 app, and conducting secondary research and conducted competitor analysis.
Competitor Analysis
We studied both our direct and indirect competitors, used their apps, mapped their user flow, their USP and their target audience. We also researched on online forums to understand what was working and what not in these apps
Major Insights
The Market is Split Between "Control" and "Convenience"
We found that apps either gave users full control to customize every detail or total convenience by letting AI build the team quickly. There wasn't much in between.
A "Special Hook" is Key
The most successful apps had a unique feature to keep users coming back, like a special game, advanced tools, or letting users copy teams to other apps
No One Explained "Why"
This was our biggest finding. Other apps used AI to suggest teams, but none of them explained why certain players were chosen. This was a huge opportunity for us to build trust by making our AI transparent.
Chapter 3
Insights to Product Strategy
We synthesized direct insights from 6 in-depth interviews and a 50+ user survey into a clear product strategy built on archetypes, competitor gaps, and user journeys.
We translated our interview and survey findings into three core User Archetypes. This helped us spot a key market gap in transparent AI during our Competitive Analysis. Finally, we mapped each archetype's journey to pinpoint their frustrations and our biggest design opportunities.
User Interview Insights
Users Aren't All the Same
Users Aren't All the Same
We found three distinct user types: Beginners (need guidance, avoid risk), Casual Players (want a quick, fun experience), and Strategic Experts (do their own research, seek high-risk/high-reward contests).
No Trust in AI without transparency
No Trust in AI without transparency
Users will not risk money on an AI-built team unless they get a clear, simple explanation for why each player was selected.
Team Creation is Too Slow and Scattered
Team Creation is Too Slow and Scattered
All users found the process inefficient. Experts hate juggling multiple apps for data, while novices feel overwhelmed. Everyone wants a faster, all-in-one solution.
Experts Use "Risk" to Create "Unique" Teams
Experts Use "Risk" to Create "Unique" Teams
For experts, a "risky" pick is a calculated strategy. They intentionally choose less-popular players to build a unique lineup, which is key to winning large tournaments.
User Survey Insights (based on 57 responses)
84%
84%
want a transparent AI-based team recommendation platform
70.2 %
70.2 %
spend >10 minutes creating a fantasy cricket team
72.5%
72.5%
want "quick" comparison between players
Who are our users?
From our research, we segmented users into 3 groups. These classifications were driven by key factors: their gameplay strategy, risk appetite, sports knowledge, time spent on the app, and their reasons for playing.
Mapping the flow
We mapped the current, frustrating journey for each archetype. This visual exercise made their pain points tangible and revealed the exact opportunities for our design to make an impact.
Chapter 4
Opportunity to stand out
Based on our research, we pinpointed the core problems driving all other user issues. We then reframed these problems as clear design opportunities and formulated our guiding "How Might We" statements.
One-Size-Fits-All
One-Size-Fits-All
How might we design a flexible experience that offers simple guidance for beginners while providing the deep data and control that experts demand?
How might we design a flexible experience that offers simple guidance for beginners while providing the deep data and control that experts demand?
Lack of Trust
Lack of Trust
How might we build user trust by designing an interface that makes AI recommendations simple, transparent, and easy to understand?
How might we build user trust by designing an interface that makes AI recommendations simple, transparent, and easy to understand?
Scattered Data
Scattered Data
How might we create a single, efficient interface that integrates all the essential stats and data users need to make a confident decision?
How might we create a single, efficient interface that integrates all the essential stats and data users need to make a confident decision?
High-Risk Strat.
High-Risk Strategy
How might we provide tools for experts to build unique lineups and a practice arena to test their high-risk strategies safely?
How might we provide tools for experts to build unique lineups and a practice arena to test their high-risk strategies safely?
Chapter 5
Building the Blueprint
We translated our team's whiteboarding brainstorm into a formal blueprint, defining our core features, Information Architecture, and user flows.
With our "How Might We" statements as our guide, we brought the entire team together—including our ML and dev specialists—for a collaborative whiteboarding session. This was a high-energy phase where we brainstormed a wide array of potential features to solve our users' core problems.
Chapter 6
The Ideation Phase
We moved from broad ideation to low-fidelity wireframes to lock the structure, then selected our visual language, giving us a clear path to the final high-fidelity design.
We held several whiteboarding sessions to brainstorm dozens of different UI approaches to solve our users' core problems—especially how to best visualize "trust" and "transparency."
Chapter 7
The Final Solution
Our final solution is an AI-powered co-pilot that provides a simple, transparent, and customizable team-building experience for all three user archetypes.
Our design team translated our wireframes and user flows into a polished, high-fidelity UI. Here is how our key features solved the core user problems:
Homepage
The homepage is a central dashboard that shows live scores and upcoming matches, with community posts and the latest news on the side to keep users engaged and informed.
Video Preview
Beginner Flow
This screen starts the team-building process for beginners, allowing them to view lineups first. They can use the simple, toggled-on "AI-Recommended Team" option to instantly generate a squad with one click.
Video Preview
Advance Flow
The "Advanced" flow gives pro players full control, to set various parameters includeing player/team balance, spin/seam balance and access to AI-driven "Optimal Composition" insights to build a highly customized, strategic squad.
Video Preview
All Your Stats in One Place
An option to view stats is readily available. users no longer need to navigate to other apps or websites for articles, news or stats. match specific stats are only a click away at Dearmteam.
Video Preview
Transparent AI Picks
We show the simple, data-driven reason for every AI-selected player. This builds trust by making the AI transparent and helps new users learn effective team-building strategies.
Video Preview
Chapter 7
The Impact
We won the Bronze Medal, beating 23 IITs, because we were the only team that truly impressed the judges with our deep user research and human-centric design.
With our final, working solution, we traveled to IIT Bombay to present. On the day of the competition, we faced an unexpected last-minute hiccup with our ML model. We had to lead with our strength: our deep, user-centric design process. We told the judges the story of our users, our archetypes, and how every feature was a direct answer to a real user need.
Final Chapter
Retrospective
My biggest takeaway was that a deep, user-centric research process isn't just a step in the project—it's the foundation for the entire solution
This high-stakes project was a rapid lesson in leadership, collaboration, and the power of a user-first mindset.
Key Learnings
Cross-Functional Collaboration is Key:
This project was an intense, real-world lesson in working with ML and development teams. I learned how to communicate design decisions, navigate technical constraints, and find solutions under intense pressure.
Research is Your Strongest Asset:
When our tech faced a last-minute hiccup, our deep understanding of the user became our most defensible point, proving to the judges that empathy is the true foundation of innovation.
Leading a Design Team:
Leading a team of 6 through this intense competition was a powerful lesson in mentorship, delegating tasks, and maintaining a clear, unified vision under pressure.