Natasha AI
A voice-enabled solution for someone who wants to design and build an app.
Listening
Design System
Creating a unified design system to bring consistency across Builder.ai products, improving efficiency in design tasks by 80%
Other Work
Builder Studio
Builder Studio
A flagship platform that bridges the gap between customers and development teams. Powered by AI-driven conversations and machine learning–based feature dependency mapping, it simplifies requirement gathering and helps users shape their app ideas into reality.
A flagship platform that bridges the gap between customers and development teams. Powered by AI-driven conversations and machine learning–based feature dependency mapping, it simplifies requirement gathering and helps users shape their app ideas into reality.
Status
Shipped
Timeline
2023 — 2025 (Over multiple phases)
Tools
Mixpanel • HotJar • Figma • Figjam • Veed.io • After effects
Team
6+
My Role
User Testing • Brainstorming • Competitor Analysis • Systems Thinking • Conversational Design • Wireframing • Workshops • Prototyping
Status
Shipped
Timeline
2023 — 2025 (On mulitple phases)
Tools
Mixpanel • HotJar • Figma • Figjam • Veed.io • After effects
Team
6+
My Role
User Testing • Brainstorming • Competitor Analysis • Systems Thinking • Conversational Design • Wireframing • Workshops • Prototyping
About Builder Studio
Builder Studio 3.0 is Builder’s flagship platform, an AI-powered tool designed to help anyone bring their app idea to life. It enables users to effortlessly drag and drop features, set delivery timelines, and customise their app with simplicity and speed.
The Problem
The older version of Builder.ai’s Studio 3.0 lets users drag and drop different feature blocks to create the app they want. Oftentimes, users can choose a pre-existing popular app and take inspiration from its feature list. However, this approach has a problem: it is very limiting and not flexible at all. It often fails to collect granular but crucial information for the user to create a job requirement board.
Builder.ai needed a more proficient way to allow users to provide their requirements in a more structured manner, along with offering more information about their delivery preferences and details of their app.
The Goal
The goal was to enhance buildcard quality by capturing richer, more relevant data, helping users make clearer decisions. This aimed to reduce cart abandonment and, in turn, boost overall conversion rates.
Impact & Results
50% more data captured enriched buildcard inputs led to significantly improved data quality.
28% reduction in cart abandonment streamlined flow encouraged users to stay and complete their build.
15% improvement in overall conversion enhanced usability and flow translated directly into better conversion rates.
Highlight Videos
Studio 4 - AI powered app creator
Initial release I worked on - focused on guided app creation using AI.
Highlight Videos
Studio 4 - AI powered app creator
Initial release I worked on - focused on guided app creation using AI.
Highlight Videos
Studio 4 - AI powered app creator
Initial release I worked on - focused on guided app creation using AI.
Highlight Videos
Studio 4 - AI powered app creator
Initial release I worked on - focused on guided app creation using AI.
Studio 5 - The future of Natasha AI
Next iteration of Studio 4
Studio 5 - Latest evolution
Voice Interaction and refinement
Enjoyed the overview? Buckle up, its a long one for the curious minds.
About Builder Studio 3.0
Builder Studio 3.0 is Builder’s flagship platform, an AI-powered tool designed to help anyone bring their app idea to life. It enables users to effortlessly drag and drop features, set delivery timelines, and customise their app with simplicity and speed.
Discovery
Audit: Evaluating the Current Builder Studio Experience
We began with a heuristic evaluation using Nielsen’s heuristics, involving a team of myself and other product designers, who independently reviewed the interface. By prioritising issues by severity, we proposed actionable solutions to improve the overall user experience. The benefits of this evaluation include early detection of usability problems, cost-effective analysis, and ensuring an intuitive, user-friendly product.
The Problem
Audit: Evaluating the Current Builder Studio Experience
We began with a heuristic evaluation using Nielsen’s heuristics, involving a team of myself and other product designers, who independently reviewed the interface. By prioritising issues by severity, we proposed actionable solutions to improve the overall user experience. The benefits of this evaluation include early detection of usability problems, cost-effective analysis, and ensuring an intuitive, user-friendly product.
Understanding Current Users
To understand how users were interacting with Builder Studio’s existing flow, we combined quantitative data from Hotjar and Mixpanel with qualitative insights from sales and support call recordings. Hotjar revealed key usability issues through heatmaps and session recordings:
Scroll maps showed that 72% of users never reached the feature library, which sat below the fold, explaining low feature selection rates.
Session recordings highlighted hesitation and erratic mouse movements around the delivery stage, often followed by exits or long idle times, signs of user confusion.
Click heatmaps showed high interaction with inactive or non-clickable UI elements, especially on the preview screen, indicating misaligned expectations.
Sales and support call recordings revealed that many users struggled to articulate their app idea without clear guidance and often didn’t understand what was expected of them during key steps in the journey. Several users voiced frustration with Natasha, describing her as more of a scripted chatbot than a responsive AI, unable to adapt to their input or provide meaningful help.
The older version of Builder.ai’s Studio 3.0 lets users drag and drop different feature blocks to create the app they want. Oftentimes, users can choose a pre-existing popular app and take inspiration from its feature list. However, this approach has a problem: it is very limiting and not flexible at all. It often fails to collect granular but crucial information for the user to create a job requirement board.
Builder.ai needed a more proficient way to allow users to provide their requirements in a more structured manner, along with offering more information about their delivery preferences and details of their app.
Enhancing Buildcard quality by capturing richer data
Providing a clear journey to completion
Reducing cart abandonment
Increasing Buildcard completion rate
Boosting overall conversion rates
User Interviews
We conducted interviews with startup founders and product managers actively exploring no-code tools. While their goals varied, pain points started to surface.
User Interviews - Studio 3.0
What we found out
Drag-and-Drop Limitations
While Studio 3.0’s block-based builder simplified getting started faster, it was extremely manual and slow, and it didn’t capture crucial context
Template Over-Reliance
Users gravitated toward popular app templates (e.g, Uber clones) but these often misaligned with unique business needs.
Feature Organisation Chaos
Without role-based grouping, features became cluttered, making it hard for developers to inperpret scope.
Contextual Gaps
Users lacked the ability to annotate features with notes or specify intent with detailed information further hampering clarity.
Delivery Confusion
Ambiguous choices around timeline, platforms, and pricing created hesitation and suprise costs at the end.
Cognitive Overload
The multi-step process, through necessary, risked overwhelming users without progressive guidance.
Founders and product managers seek low-code/no-code solutions with clear guidance
Many of these users aren’t technical, so they want tools that let them build or spec an app without coding, but they still need help understanding what to do next or how to express their idea clearly. They don’t just want tools, they want smart guidance built into the process.
Traditional app briefs lack structure, often miss critical details
When users describe their app idea (in a form or during onboarding), they often leave out important information, like who the users are, what features they actually need, or how it should work. Their briefs are vague, which causes delays or miscommunication in the build process.
Inspiration & Competitor Analysis
Our competitive analysis highlighted a clear gap in the market. While tools like Bubble and AppyPie offer strong drag-and-drop functionality, they rely on users to manually define their feature list. For those still in the ideation phase, and needed guidance to translate their business concepts into technical requirements. This approach allowed us to differentiate meaningfully and secure a unique place in the market.
We benchmarked key no-code and low-code platforms, like Appy Pie, Bubble.io, and Webflow - focusing on three core areas: sign-up experience, self-serve guidance, and human interaction. Our evaluation revealed that while many competitors offered easy sign-up and rich educational content, most lacked proactive human support and relied heavily on users to figure things out alone.
Live chat was often slow or robotic. For example, on Appy Pie, it took over 30 minutes to get a useful response, despite the promise of 24/7 support. Most platforms offered documentation but missed opportunities to guide users contextually within the product.
These insights showed a clear opportunity for Builder to differentiate with structured onboarding, embedded AI support, and fast, contextual human help, especially for first-time founders unfamiliar with technical terminology.
User Archetypes
We focused our strategy on The Executor. A high-intent, ROI-driven user most likely to convert. Their need for clarity, structure, and trust informed key design decisions around onboarding, pricing transparency, and delivery, helping reduce friction at critical conversion points.
Design Challenges & Goals
Our central question began to form.
How might we guide users to articulate detailed requirements, within an intuitive, structured flow?
We prioritised three end goals:
Streamlined Requirement Capture
AI-guided wizard translating user ideas into actionable modules and journeys.
Intelligent Feature Planning
Role-based grouping of features, complete with granular metadata and real time previews.
Transparent Delivery Process
Progressive checkout with live cost/timeline estimates and clear ownership/export options.
Design Challenges & Goals
How might we guide users to articulate detailed requirements, within an intuitive, structured flow?
Workshops
To better understand user motivations and breakdowns across the journey, we ran collaborative workshops with the UX research team focused on two areas: user compromises and “wow” moments, which highlighted friction and delight points.
Key Compromises Uncovered
Lack of Clarity & Guidance
This directly affected user confidence and completion. Internal jargon confused users, and they weren’t clear on what they were building or getting. It’s foundational and likely caused drop-off.
Weak Value Communication
If users didn’t understand the platform’s value or what made it magical, they were less likely to stay engaged or convert. This touches both UX and marketing.
Poor Timing of Visualisation & Feedback
Delaying previews and intelligent guidance meant missed chances to excite users early — a big issue for first impressions and momentum.
WOW Moments Identified
Clear Visual Previews
When previews matched what users picked, it helped them feel in control and understand what they were building
Support that feels human
Timely, friendly messages (from AI or humans) made users feel guided and reassured — especially during tricky steps.
User Journey flows
We mapped the user journey to uncover key friction points and opportunities. This helped us identify where users felt confused or unsupported, and informed where to focus our design efforts.
Designing the System & Flows
Agentic Structure: Translating Ideas into Buildable Outputs
Based on early research, we collaborated with the data science team to define a technical structure that Natasha could use to interpret vague user input. We scoped what was achievable with machine learning and used that to design a flow that turns open-ended ideas into structured, buildable outputs. For example, when a user says, “I want to build an app like Uber,” Natasha follows a logic model to ask what platform it’s for, who the users are, and what core journeys and features are needed. This ensured that even high-level ideas could be accurately scoped into Buildcards.
Diagnosing Pain Points Along the Journey
Beyond mapping the user flow, we broke it down to expose user pain points and Natasha’s supporting role at each step. This layered view helped us fine-tune the assistant’s prompts, surfacing where users get stuck and how to guide them toward clearer outcomes.
Conversational Design
To shape Natasha’s voice and guide the user journey, our copy team collaborated with us closely on the conversational design. We mapped real example dialogues, tested tone and phrasing, and refined each prompt to feel intuitive and human, helping users feel supported, not scripted.
Solidifying Core Stages
To shape Natasha’s voice and guide the user journey, our copy team collaborated with us closely on the conversational design. We mapped real example dialogues, tested tone and phrasing, and refined each prompt to feel intuitive and human, helping users feel supported, not scripted.
Conceptualisation - Wireframing
I worked with the content and copy team to turn abstract concepts into clear, focused layouts. Using research and journey mapping, we shaped the app’s structure to feel intuitive and easy to follow. The wireframes outlined key screens and content, and we iterated on them to better match how users actually think and navigate.
Solution Highlights
Introducing A Better Concept
To help users better articulate what they needed, I introduced a new nested structure: user types → journeys → features. Rather than starting from a long list of technical features, users could now describe their app based on who it’s for and what it needs to do. This shift made the process more intuitive and collaborative — users felt less overwhelmed, and the team could align more easily on scope. It also laid the groundwork for a smarter, more flexible way to recommend and organise features through journeys.
Guided onboarding
A seamless, step-by-step onboarding process with clear prompts and tooltips. This guided journey empowers users to quickly master the app with confidence.
Reducing choice of paradox
In the older version, users saw too many unrelated, preset app options, causing overwhelm and hesitation.
The redesign brings tailored, relevant features upfront, making the process feel clearer and helping users decide faster with more confidence.
A collection of 500+ known apps
Referencing familiar apps helps them communicate ideas more clearly and effectively, speeding up the design process and improving collaboration. This ensures the final product feels intuitive and grounded in real-world use cases.
Users can use publicly available apps as inspiration for their own projects.
Quick Visualisation
Users get a clear view of their app upfront, with features grouped by user role and a matching splash screen generated by AI. This helps them align on the concept early, reduces misunderstandings, and builds confidence to move forward.
Customising Look & Feel
Users can personalise their app experience by adjusting their app’s appearance and functionality during this stage.
Users can upload their logo, and an appropriate colour will be automatically selected. They will also have options to modify various visual aspects of the app’s look and feel, such as buttons, fonts, colours, and more.
Guidance for new user
Many of our users aren’t highly technical, so they rely heavily on the interactive guidance to navigate the process with confidence.
Customising App Requirements
The Refine Idea stage allows users to fine-tune their app’s requirements.
Users can review all the user roles, journeys, and associated features recommended by our AI feature generator. They can then decide whether to add or remove any specific feature or journey.
Deciding the Delivery Preferences
Users gain precise control over their platform, customising it to match their vision. They can set their own development pace and choose how to deploy their app in the cloud, creating a tailored, seamless experience aligned with their goals and timeline.
Final Review
A final review screen lets users confirm all app details and interact with a live prototype. They can easily share it with others or book a call with an expert if they need additional support.
Exploration of AI generated splash screen
We also designed over 1000+ mobile and web screen templates to cover various verticals, ensuring consistency and scalability. This allowed us to construct UI for our prototypes and app visualisation.
Studio 5.0
After launching Studio 4.0 in 2024, our design team initiated an R&D effort to reimagine the platform, dubbed Studio 5.0. Inspired by the demand and rise of voice interface in popular application like ChatGPT and Gemini AI, we prioritised voice interaction to create a hands-free experience, significantly speeding up user input collection. We also enhanced AI capabilities, focusing on:
- Improved conversational AI-generated mood boards for visual direction
- Smarter requirement gathering.
- Refined idea iteration to better align with user goals.
This was also an opportunity to modernise the platform’s look and feel. A key challenge was ensuring cohesion across our evolving product suite, which led us to develop an upgraded Builder Design Language, Block 5.0, a more intuitive, scalable, and visually unified design system.
Results & Impact
Understanding the measurable outcomes and key insights gained throughout the design process has been crucial. It highlights how our approach shaped the user experience and uncovered future opportunities for growth. This reflection also helps me grow as a designer, empowering me to create more intuitive and user-centric flows in future projects.
50% more data captured
Enriched buildcard inputs led to significantly improved data quality.
70% increase in journey completion
More users successfully reached the end of the experience.
28% reduction in cart abandonment
Streamlined flow encouraged users to stay and complete their build.
125% uplift in buildcard completion rate
Major progress in engagement with the core feature.
15% improvement in overall conversion
Enhanced usability and flow translated directly into better conversion rates.
4.6/5
NPS score
Marked improvement in user satisfaction and experience.
Next steps
Integrate analytics to monitor drop-off, build AI-powered feature-grouping recommendations, and introduce real-time collaboration for journey co-editing.
Next up
Builder Cloud
Keeping track of your cloud usage, a Dashboard re-design.
UX UI / Dashboard
Final Designs
Conversational AI agent (Natasha)
Captures requirements through text or voice, suggests modules and features, and asks smart follow-ups automatically.
Feature Library
A wide range of journeys and features with searchable tags, real-world inspiration, and customisation options.
Progressive Checkout
Checklist-style flow with live pricing and timeline updates. Users can save progress and return later.
Timeline
July 2024 - December 2024
Project Overview
Project Overview
Timeline
2023 — 2025 (Over multiple phases)
Timeline
July 2024 - December 2024
Tools
Mixpanel • HotJar • Figma • Figjam • Veed.io • After effects
Team
6+
My Role
User Testing • Brainstorming • Competitor Analysis • Systems Thinking • Conversational Design • Wireframing • Workshops • Prototyping
About Builder Studio
About Builder Studio
Builder Studio 3.0 is Builder’s flagship platform, an AI-powered tool designed to help anyone bring their app idea to life. It enables users to effortlessly drag and drop features, set delivery timelines, and customise their app with simplicity and speed.
The Problem
The Problem
The older version of Builder.ai’s Studio 3.0 lets users drag and drop different feature blocks to create the app they want. Oftentimes, users can choose a pre-existing popular app and take inspiration from its feature list. However, this approach has a problem: it is very limiting and not flexible at all. It often fails to collect granular but crucial information for the user to create a job requirement board.
Builder Now is seeing a significant drop-off rate of 92% when users reach the current platform. Users struggled to visualise their app ideas in a clear and simple way, making it harder for them to showcase their concept to investors, gain confidence in their vision, and fully understand their app idea before committing to a significant investment with Builder.ai.
Builder.ai needed a more proficient way to allow users to provide their requirements in a more structured manner, along with offering more information about their delivery preferences and details of their app.
Builder Now is seeing a significant drop-off rate of 92% when users reach the current platform. Users struggled to visualise their app ideas in a clear and simple way, making it harder for them to showcase their concept to investors, gain confidence in their vision, and fully understand their app idea before committing to a significant investment with Builder.ai.
The Goal
The Goal
The goal was to enhance buildcard quality by capturing richer, more relevant data, helping users make clearer decisions. This aimed to reduce cart abandonment and, in turn, boost overall conversion rates.
Builder Now is seeing a significant drop-off rate of 92% when users reach the current platform. Users struggled to visualise their app ideas in a clear and simple way, making it harder for them to showcase their concept to investors, gain confidence in their vision, and fully understand their app idea before committing to a significant investment with Builder.ai.
Impact & Results
Impact & Results
50% more data captured enriched buildcard inputs led to significantly improved data quality.
28% reduction in cart abandonment streamlined flow encouraged users to stay and complete their build.
15% improvement in overall conversion enhanced usability and flow translated directly into better conversion rates.
Final Designs
Final Designs
Guided onboarding
A seamless, step-by-step onboarding process with clear prompts and tooltips. This guided journey empowers users to quickly master the app with confidence.
Reducing choice of paradox
In the older version, users saw too many unrelated, preset app options, causing overwhelm and hesitation.
The redesign brings tailored, relevant features upfront, making the process feel clearer and helping users decide faster with more confidence.
A collection of 500+ known apps
Users can use publicly available apps as inspiration for their own projects.
Referencing familiar apps helps them communicate ideas more clearly and effectively, speeding up the design process and improving collaboration. This ensures the final product feels intuitive and grounded in real-world use cases.
Quick Visualisation
Users get a clear view of their app upfront, with features grouped by user role and a matching splash screen generated by AI. This helps them align on the concept early, reduces misunderstandings, and builds confidence to move forward.
Guidance for new user
Many of our users aren’t highly technical, so they rely heavily on the interactive guidance to navigate the process with confidence.
Customising Look & Feel
Users can personalise their app experience by adjusting their app’s appearance and functionality during this stage.
Users can upload their logo, and an appropriate colour will be automatically selected. They will also have options to modify various visual aspects of the app’s look and feel, such as buttons, fonts, colours, and more.
Customising App Requirements
The Refine Idea stage allows users to fine-tune their app’s requirements.
Users can review all the user roles, journeys, and associated features recommended by our AI feature generator. They can then decide whether to add or remove any specific feature or journey.
Journey Library
The journey library offers a collection of common, pre-built app journeys, each bundled with the key features needed to complete it. Users can search, review recommendations, and easily add relevant journeys to their app.
Reducing choice of paradox
In the older version, users saw too many unrelated, preset app options, causing overwhelm and hesitation.
The redesign brings tailored, relevant features upfront, making the process feel clearer and helping users decide faster with more confidence.
Feature Library
The feature library contains a range of commonly used, pre-built app features. Users can browse recommendations or search and add features directly to their app with ease.
Detailed Journey Features
Each journey or feature includes clear, detailed information upfront to help users decide if it suits their needs. A short video walkthrough is also available for a quick overview.
Deciding the Delivery Preferences
Users gain precise control over their platform, customising it to match their vision. They can set their own development pace and choose how to deploy their app in the cloud, creating a tailored, seamless experience aligned with their goals and timeline.
Final Review
A final review screen lets users confirm all app details and interact with a live prototype. They can easily share it with others or book a call with an expert if they need additional support.
Highlight Videos
Highlight Videos
Studio 4 - AI powered app creator
Initial release I worked on - focused on guided app creation using AI.
Studio 5 - The future of Natasha AI
Next iteration of Studio 4
Studio 5 - Latest evolution
Voice Interaction and refinement
Enjoyed the overview? Buckle up, its a long one for the curious minds
Extended case study on Notion.
About Builder Studio 3.0
About Builder Studio 3.0
Builder Studio 3.0 is Builder’s flagship platform, an AI-powered tool designed to help anyone bring their app idea to life. It enables users to effortlessly drag and drop features, set delivery timelines, and customise their app with simplicity and speed.
Studio 3.0 - Flagship Platform
Studio 3.0 - Flagship Platform
Discovery
Discovery
Audit: Evaluating the Current Builder Studio Experience
We began with a heuristic evaluation using Nielsen’s heuristics, involving a team of myself and other product designers, who independently reviewed the interface. By prioritising issues by severity, we proposed actionable solutions to improve the overall user experience. The benefits of this evaluation include early detection of usability problems, cost-effective analysis, and ensuring an intuitive, user-friendly product.
The Problem
The Problem
The older version of Builder.ai’s Studio 3.0 lets users drag and drop different feature blocks to create the app they want. Oftentimes, users can choose a pre-existing popular app and take inspiration from its feature list. However, this approach has a problem: it is very limiting and not flexible at all. It often fails to collect granular but crucial information for the user to create a job requirement board.
Builder.ai needed a more proficient way to allow users to provide their requirements in a more structured manner, along with offering more information about their delivery preferences and details of their app.
Enhancing Buildcard quality by capturing richer data
Providing a clear journey to completion
Reducing cart abandonment
Increasing Buildcard completion rate
Boosting overall conversion rates
Understanding Current Users
What we knew already
Internal analytics showed the dashboard was underperforming compared to the rest of the product.
When reviewing Mixpanel, we found:
35% of users dropped off shortly after landing on the dashboard — suggesting it wasn’t clear or useful enough to keep them engaged.
It had only a 20% retention rate, meaning few users came back to use it again.
From support tickets, we also learned:
Many users experienced unexpected service suspensions after forgetting to top up their wallet.
These incidents were often caused by poor visibility of the Auto-Recharge feature, which was buried in the UI and rarely discovered until it was too late.
But what else was stopping users from using the dashboard?
To dig deeper, we conducted customer interviews.
Customer interviews
In-depth interviews were conducted with 5 existing Builder.ai customers who had previously used the "Cloud" platform to track their spend. These interviews focused on understanding their experiences, challenges, and expectations for using the current cloud dashboard.
To understand how users were interacting with Builder Studio’s existing flow, we combined quantitative data from Hotjar and Mixpanel with qualitative insights from sales and support call recordings. Hotjar revealed key usability issues through heatmaps and session recordings:
What we knew already
Internal analytics showed the dashboard was underperforming compared to the rest of the product.
When reviewing Mixpanel, we found:
35% of users dropped off shortly after landing on the dashboard — suggesting it wasn’t clear or useful enough to keep them engaged.
It had only a 20% retention rate, meaning few users came back to use it again.
From support tickets, we also learned:
Many users experienced unexpected service suspensions after forgetting to top up their wallet.
These incidents were often caused by poor visibility of the Auto-Recharge feature, which was buried in the UI and rarely discovered until it was too late.
But what else was stopping users from using the dashboard?
To dig deeper, we conducted customer interviews.
Customer interviews
In-depth interviews were conducted with 5 existing Builder.ai customers who had previously used the "Cloud" platform to track their spend. These interviews focused on understanding their experiences, challenges, and expectations for using the current cloud dashboard.
Scroll maps showed that 72% of users never reached the feature library, which sat below the fold, explaining low feature selection rates.
Session recordings highlighted hesitation and erratic mouse movements around the delivery stage, often followed by exits or long idle times, signs of user confusion.
Click heatmaps showed high interaction with inactive or non-clickable UI elements, especially on the preview screen, indicating misaligned expectations.
Sales and support call recordings revealed that many users struggled to articulate their app idea without clear guidance and often didn’t understand what was expected of them during key steps in the journey. Several users voiced frustration with Natasha, describing her as more of a scripted chatbot than a responsive AI, unable to adapt to their input or provide meaningful help.
What we knew already
Internal analytics showed the dashboard was underperforming compared to the rest of the product.
When reviewing Mixpanel, we found:
35% of users dropped off shortly after landing on the dashboard — suggesting it wasn’t clear or useful enough to keep them engaged.
It had only a 20% retention rate, meaning few users came back to use it again.
From support tickets, we also learned:
Many users experienced unexpected service suspensions after forgetting to top up their wallet.
These incidents were often caused by poor visibility of the Auto-Recharge feature, which was buried in the UI and rarely discovered until it was too late.
But what else was stopping users from using the dashboard?
To dig deeper, we conducted customer interviews.
Customer interviews
In-depth interviews were conducted with 5 existing Builder.ai customers who had previously used the "Cloud" platform to track their spend. These interviews focused on understanding their experiences, challenges, and expectations for using the current cloud dashboard.
User Interviews
User Interviews
We conducted interviews with startup founders and product managers actively exploring no-code tools. While their goals varied, pain points started to surface.
User Interviews - Studio 3.0
User Interviews - Studio 3.0
What we found out
What we found out
Drag-and-Drop Limitations
While Studio 3.0’s block-based builder simplified getting started faster, it was extremely manual and slow, and it didn’t capture crucial context
Template Over-Reliance
Users gravitated toward popular app templates (e.g, Uber clones) but these often misaligned with unique business needs.
Feature Organisation Chaos
Without role-based grouping, features became cluttered, making it hard for developers to inperpret scope.
Contextual Gaps
Users lacked the ability to annotate features with notes or specify intent with detailed information further hampering clarity.
Delivery Confusion
Ambiguous choices around timeline, platforms, and pricing created hesitation and suprise costs at the end.
Cognitive Overload
The multi-step process, through necessary, risked overwhelming users without progressive guidance.
Founders and product managers seek low-code/no-code solutions with clear guidance
→ Many of these users aren’t technical, so they want tools that let them build or spec an app without coding, but they still need help understanding what to do next or how to express their idea clearly. They don’t just want tools, they want smart guidance built into the process.
Traditional app briefs lack structure, often miss critical details
→ When users describe their app idea (in a form or during onboarding), they often leave out important information—like who the users are, what features they actually need, or how it should work. Their briefs are vague, which causes delays or miscommunication in the build process.
Inspiration & Competitor Analysis
Our competitive analysis highlighted a clear gap in the market. While tools like Bubble and AppyPie offer strong drag-and-drop functionality, they rely on users to manually define their feature list. For those still in the ideation phase, and needed guidance to translate their business concepts into technical requirements. This approach allowed us to differentiate meaningfully and secure a unique place in the market.
We benchmarked key no-code and low-code platforms, like Appy Pie, Bubble.io, and Webflow - focusing on three core areas: sign-up experience, self-serve guidance, and human interaction. Our evaluation revealed that while many competitors offered easy sign-up and rich educational content, most lacked proactive human support and relied heavily on users to figure things out alone.
Live chat was often slow or robotic. For example, on Appy Pie, it took over 30 minutes to get a useful response, despite the promise of 24/7 support. Most platforms offered documentation but missed opportunities to guide users contextually within the product.
These insights showed a clear opportunity for Builder to differentiate with structured onboarding, embedded AI support, and fast, contextual human help, especially for first-time founders unfamiliar with technical terminology.
The Goal
Our goal was to redesign the dashboard to make spend tracking clearer, improve discoverability of Auto-Recharge, and reduce reliance on customer support—ultimately driving better engagement and retention.
User Archetypes
User Archetypes
We focused our strategy on The Executor. A high-intent, ROI-driven user most likely to convert. Their need for clarity, structure, and trust informed key design decisions around onboarding, pricing transparency, and delivery, helping reduce friction at critical conversion points.
Design Challenges & Goals
Our central question began to form.
How might we guide users to articulate detailed requirements, within an intuitive, structured flow?
We prioritised three end goals:
Streamlined Requirement Capture
AI-guided wizard translating user ideas into actionable modules and journeys.
Intelligent Feature Planning
Role-based grouping of features, complete with granular metadata and real time previews.
Transparent Delivery Process
Progressive checkout with live cost/timeline estimates and clear ownership/export options.
Workshops
Workshops
To better understand user motivations and breakdowns across the journey, we ran collaborative workshops with the UX research team focused on two areas: user compromises and “wow” moments, which highlighted friction and delight points.
Key Compromises Uncovered
Key Compromises Uncovered
Lack of Clarity & Guidance
This directly affected user confidence and completion. Internal jargon confused users, and they weren’t clear on what they were building or getting. It’s foundational and likely caused drop-off.
Weak Value Communication
If users didn’t understand the platform’s value or what made it magical, they were less likely to stay engaged or convert. This touches both UX and marketing.
Poor Timing of Visualisation & Feedback
Delaying previews and intelligent guidance meant missed chances to excite users early — a big issue for first impressions and momentum.
WOW Moments Identified
WOW Moments Identified
Clear Visual Previews
When previews matched what users picked, it helped them feel in control and understand what they were building
Support that feels human
Timely, friendly messages (from AI or humans) made users feel guided and reassured — especially during tricky steps.
User Journey flows
User Journey flows
We mapped the user journey to uncover key friction points and opportunities. This helped us identify where users felt confused or unsupported, and informed where to focus our design efforts.
Designing the System & Flows
Agentic Structure: Translating Ideas into Buildable Outputs
Based on early research, we collaborated with the data science team to define a technical structure that Natasha could use to interpret vague user input. We scoped what was achievable with machine learning and used that to design a flow that turns open-ended ideas into structured, buildable outputs. For example, when a user says, “I want to build an app like Uber,” Natasha follows a logic model to ask what platform it’s for, who the users are, and what core journeys and features are needed. This ensured that even high-level ideas could be accurately scoped into Buildcards.
Diagnosing Pain Points Along the Journey
Beyond mapping the user flow, we broke it down to expose user pain points and Natasha’s supporting role at each step. This layered view helped us fine-tune the assistant’s prompts, surfacing where users get stuck and how to guide them toward clearer outcomes.
Conversational Design
To shape Natasha’s voice and guide the user journey, our copy team collaborated with us closely on the conversational design. We mapped real example dialogues, tested tone and phrasing, and refined each prompt to feel intuitive and human, helping users feel supported, not scripted.
Solidifying Core Stages
We broke the app-building process into clear, structured stages, from initial AI chat to final sign-off. At each step, we refined the experience to improve clarity, reduce user friction, and ensure smooth handoffs. This visual flow helped align teams and build a more intuitive journey for users.
Conceptualisation - Wireframing
Conceptualisation - Wireframing
I worked with the content and copy team to turn abstract concepts into clear, focused layouts. Using research and journey mapping, we shaped the app’s structure to feel intuitive and easy to follow. The wireframes outlined key screens and content, and we iterated on them to better match how users actually think and navigate.
Solution Highlights
Introducing A Better Concept
To help users better articulate what they needed, I introduced a new nested structure: user types → journeys → features. Rather than starting from a long list of technical features, users could now describe their app based on who it’s for and what it needs to do. This shift made the process more intuitive and collaborative — users felt less overwhelmed, and the team could align more easily on scope. It also laid the groundwork for a smarter, more flexible way to recommend and organise features through journeys.
Nested concept of user types > journeys > features
Conversational AI agent (Natasha)
Captures requirements through text or voice, suggests modules and features, and asks smart follow-ups automatically.
Feature Library
A wide range of journeys and features with searchable tags, real-world inspiration, and customisation options.
Progressive Checkout
Checklist-style flow with live pricing and timeline updates. Users can save progress and return later.
User Testing
To validate our approach and refine the experience, we ran multiple moderated user testing sessions across a range of user types — from tech-savvy founders to non-technical first-time entrepreneurs. These sessions gave us both behavioral observations and direct user feedback, helping us assess usability, clarity, and perceived value. We found the following:
Overwhelmed by Feature Complexity
Users appreciated the ability to refine and group features, but many felt overwhelmed by the number of available options and the lack of clear guidance on prioritisation or relevance to their app type.
Lack or Clarity in the AI Chat Output
While the AI agent effectively gathered user intent, some users struggled to understand how their input translated into concrete features, leading to doubts about the accuracy and completeness of the AI-generated flow.
Uncertainity Around Expectations
The delivery configuration step raised questions about timelines, platform limitations, and pricing. Users wanted more contextual help and transparency to make informed decisions before committing to the next step.
Final Screens
Final Screens
Guided onboarding
A seamless, step-by-step onboarding process with clear prompts and tooltips. This guided journey empowers users to quickly master the app with confidence.
Reducing choice of paradox
In the older version, users saw too many unrelated, preset app options, causing overwhelm and hesitation.
The redesign brings tailored, relevant features upfront, making the process feel clearer and helping users decide faster with more confidence.
A collection of 500+ known apps
Users can use publicly available apps as inspiration for their own projects.
Referencing familiar apps helps them communicate ideas more clearly and effectively, speeding up the design process and improving collaboration. This ensures the final product feels intuitive and grounded in real-world use cases.
Quick Visualisation
Users get a clear view of their app upfront, with features grouped by user role and a matching splash screen generated by AI. This helps them align on the concept early, reduces misunderstandings, and builds confidence to move forward.
Guidance for new user
Many of our users aren’t highly technical, so they rely heavily on the interactive guidance to navigate the process with confidence.
Customising Look & Feel
Users can personalise their app experience by adjusting their app’s appearance and functionality during this stage.
Users can upload their logo, and an appropriate colour will be automatically selected. They will also have options to modify various visual aspects of the app’s look and feel, such as buttons, fonts, colours, and more.
Customising App Requirements
The Refine Idea stage allows users to fine-tune their app’s requirements.
Users can review all the user roles, journeys, and associated features recommended by our AI feature generator. They can then decide whether to add or remove any specific feature or journey.
Journey Library
The journey library offers a collection of common, pre-built app journeys, each bundled with the key features needed to complete it. Users can search, review recommendations, and easily add relevant journeys to their app.
Reducing choice of paradox
In the older version, users saw too many unrelated, preset app options, causing overwhelm and hesitation.
The redesign brings tailored, relevant features upfront, making the process feel clearer and helping users decide faster with more confidence.
Feature Library
The feature library contains a range of commonly used, pre-built app features. Users can browse recommendations or search and add features directly to their app with ease.
Detailed Journey Features
Each journey or feature includes clear, detailed information upfront to help users decide if it suits their needs. A short video walkthrough is also available for a quick overview.
Deciding the Delivery Preferences
Users gain precise control over their platform, customising it to match their vision. They can set their own development pace and choose how to deploy their app in the cloud, creating a tailored, seamless experience aligned with their goals and timeline.
Final Review
A final review screen lets users confirm all app details and interact with a live prototype. They can easily share it with others or book a call with an expert if they need additional support.
Further Improvements
Exploration of AI generated splash screen
To humanise the experience, we experimented with AI-generated splash screens using Stable Diffusion, DALL·E, and Adobe Firefly. By running 1,000+ prompts and manually curating outputs for training data, we landed on imagery that reinforced character and set the tone for each stage of the journey.
Exploration of AI generated splash screen
We also designed over 1000+ mobile and web screen templates to cover various verticals, ensuring consistency and scalability. This allowed us to construct UI for our prototypes and app visualisation.
Studio 5.0
Studio 5.0
After launching Studio 4.0 in 2024, our design team initiated an R&D effort to reimagine the platform, dubbed Studio 5.0. Inspired by the demand and rise of voice interface in popular application like ChatGPT and Gemini AI, we prioritised voice interaction to create a hands-free experience, significantly speeding up user input collection. We also enhanced AI capabilities, focusing on:
- Improved conversational AI-generated mood boards for visual direction
- Smarter requirement gathering.
- Refined idea iteration to better align with user goals.
This was also an opportunity to modernise the platform’s look and feel. A key challenge was ensuring cohesion across our evolving product suite, which led us to develop an upgraded Builder Design Language, Block 5.0, a more intuitive, scalable, and visually unified design system.
Results & Impact
Results & Impact
Understanding the measurable outcomes and key insights gained throughout the design process has been crucial. It highlights how our approach shaped the user experience and uncovered future opportunities for growth. This reflection also helps me grow as a designer, empowering me to create more intuitive and user-centric flows in future projects.
50% more data captured
Enriched buildcard inputs led to significantly improved data quality.
70% increase in journey completion
More users successfully reached the end of the experience.
28% reduction in cart abandonment
Streamlined flow encouraged users to stay and complete their build.
125% uplift in buildcard completion rate
Major progress in engagement with the core feature.
15% improvement in overall conversion
Enhanced usability and flow translated directly into better conversion rates.
4.6/5
NPS score
Marked improvement in user satisfaction and experience.
Next steps
Next steps
Integrate analytics to monitor drop-off, build AI-powered feature-grouping recommendations, and introduce real-time collaboration for journey co-editing.
Integrate analytics to monitor drop-off, build AI-powered feature-grouping recommendations, and introduce real-time collaboration for journey co-editing.
Other Work
UX UI / Dashboard
Builder Cloud
Keeping track of your cloud usage, a Dashboard re-design.
UX UI / Strategy / End to End
Builder Now
Helping Builder.ai increase conversions through users visualising their app idea..
Design System / Theming
Design System
Creating a unified design system to bring consistency across Builder.ai products.
Listening
UX UI / Interaction Design
Natasha AI: Let's Talk
A voice-enabled solution for someone who wants to design and build an app.
UX UI / Dashboard
Builder Cloud
Keeping track of your cloud usage, a Dashboard re-design.
UX UI / Strategy / End to End
Builder Now
Helping Builder.ai increase conversions through users visualising their app idea..
Design System / Theming
Design System
Creating a unified design system to bring consistency across Builder.ai products.
Listening
UX UI / Interaction Design
Natasha AI: Let's Talk
A voice-enabled solution for someone who wants to design and build an app.
UX UI / Dashboard
Builder Cloud
Keeping track of your cloud usage, a Dashboard re-design.
UX UI / Strategy / End to End
Builder Now
Helping Builder.ai increase conversions through users visualising their app idea..
Design System / Theming
Design System
Creating a unified design system to bring consistency across Builder.ai products.
Listening
UX UI / Interaction Design
Natasha AI: Let's Talk
A voice-enabled solution for someone who wants to design and build an app.
UX UI / Dashboard
Builder Cloud
Keeping track of your cloud usage, a Dashboard re-design.
UX UI / Strategy / End to End
Builder Now
Helping Builder.ai increase conversions through users visualising their app idea..
Design System / Theming
Design System
Creating a unified design system to bring consistency across Builder.ai products.
Listening
UX UI / Interaction Design
Natasha AI: Let's Talk
A voice-enabled solution for someone who wants to design and build an app.