A 5 day design sprint building a voice user interface for Indigo Airlines that enables hands-free flight booking through natural conversation.

Booking a flight on your phone requires your eyes, your hands, and your full attention. For someone rushing through an airport, managing other tasks, or visually impaired, that experience breaks down fast.

The question was not whether voice could work for flight booking. It was how to design a conversation that felt natural, handled ambiguity gracefully, and worked in the real, noisy, unpredictable conditions people actually travel in.

Four user types shaped the design direction:

 who need quick, frictionless booking without stopping what they are doing.

 who cannot afford to navigate multiple screens under time pressure.

 who rely on voice interaction for independent, accessible service and are currently underserved by screen-first booking experiences.

 managing other responsibilities simultaneously who need hands-free access to travel planning.

Designing for voice is fundamentally different from designing for screens. There are 

. Users cannot see what options exist. The interface has to guide without overwhelming, handle casual and unexpected phrasing, and recover gracefully when things go wrong.

Three specific things had to be figured out: how natural could the language be, what happens when a user says something the system does not expect, and when should the interface speak versus show.

I started with role-playing exercises, running through 3 different conversation combinations to map out natural dialogue patterns and find the path with the least cognitive load. This was about figuring out how people actually talk when they want to book a flight and designing around that.

. But the real design work happened in the edges of what if someone says "cheapest" instead of specifying a price range? What if they change their mind mid-conversation? What if they go quiet?

Every node in the dialogue flow has a fallback response. 

 states as carefully as the success path was the most important discipline of this project. A voice interface lives or dies by its recovery flows.

The interface is multimodal by design. Users can speak their choice or tap it on screen, switching between the two without losing their place in the conversation. This made the experience usable in real conditions.

 for realistic voice interaction prototyping, designed in 

 concepts for real-time flight data retrieval.

The prototype is fully interactive. Watch the full voice interaction below.

differences between voice user interfaces (VUIs) and screen-based UIs

 are crucial steps in comprehending natural language and diverse communication styles.

 in voice interactions. I realized that there can be numerous ways to express the same idea, emphasizing the importance of identifying and utilizing key phrases effectively.

This project also provided me with the opportunity to familiarize myself with a new software- 

 and developing a user flow within it. This was a departure from my usual tools and methods, offering a valuable learning experience in adapting to different platforms and workflows.

Indigo VUI: Designing a Voice Interface for Flight Booking

I have always been the kind of person who notices when something is harder than it needs to be. In high school, during the first COVID lockdown, I channelled that into building an app with a few friends to connect people with their local grocery vendors before quick commerce existed. I was essentially wireframing without knowing the word existed.
That project sent me down a rabbit hole I never came out of. Interaction design degree, jury seasons, research where I would talk to anyone willing to sit with me for twenty minutes. Turns out that is my favourite part: meeting people, hearing how they think, figuring out what would genuinely make their lives easier.
Off screen I am fuelled by good coffee, playlists I am very precious about, the gym, books I am always halfway through, losing track of time in art galleries and conversations I did not plan on having but ended up enjoying the most. 
My goal is to design for people who deserve better experiences than they usually get. Accessible, community-led, genuinely useful. That is what drives me and where I am going.

Zeron is a B2B SaaS company building GRC and vendor risk management products. I joined as a design intern and transitioned to full-time, working across both core platforms: Conformity and Vendor Pulse, designing for enterprise clients including Aditya Birla Capital, HDFC Bank, Deloitte, Angel One, and Bombay Stock Exchange.

My work spans dashboard redesigns, role-based onboarding flows, AI-integrated features, audit screens, and a cross-product task workspace; all shipped in close collaboration with product and engineering.

Digital marketing and design agency based in Bangalore. Designed landing pages and microsites for brands including Salesforce, Swiggy, Kauvery Hospital, and STG Labs; going from brief to final delivery within sprint cycles and presenting directly to clients.

O.C. Tanner is a global employee recognition platform. I audited their B2B SaaS dashboard, identified usability gaps, and delivered redesigned components with WCAG accessibility compliance.

Huminos is an OKR and performance management platform. Refined the dashboard UI, improved how goals and KPIs were structured, and put together a component style guide to bring consistency to the product's visual output.

End-to-end product design across Zeron's GRC & VRM platforms including dashboards, AI features, role-based onboarding, and audit flows.

Enterprise compliance software is typically built for feature completeness, not for how risk teams actually work. When I joined, the Vendor Pulse dashboard fragmented risk data across multiple views. Conformity onboarded every user into the same generic flow regardless of their role. There was no intelligence layer. Users were doing extra work just to use the product.

 Consolidated 6+ fragmented modules into a single unified view. CX feedback from client calls and Clarity session recordings showed users were hesitating and missing key information. The redesign was built around one question: can a risk manager scan this in under 10 seconds and know what needs attention?

Replaced a one-size-fits-all Conformity (GRC Platform) setup flow with three distinct experiences for Task Performers, Managers, and Executives, each scoped to what that role actually needs during onboarding.

 Alongside the Conformity setup flows, I designed the initial onboarding experience for new users signing onto the Zeron platform, from landing and domain confirmation through to role selection and platform setup. The screens also had to communicate Zeron's visual identity, setting the tone for the brand before users got into the product. The goal was to make the first few minutes feel guided, intentional, and distinctly Zeron.

Designed two AI features: a contextual side drawer and hover-triggered insight popovers on the Vendors table, so users get intelligence inline without leaving their current screen.

Designed both sides of the audit flow: the organisation managing assessments and the external auditor conducting them. Two distinct users, one coherent system.

A cross-product task hub pulling assigned controls, evidences, vendor assessments, and risks from all three Zeron platforms into one personalised view.

Working on GRC and vendor risk products taught me that enterprise UX lives or dies by context. The users are not exploring, they are executing, under pressure, often accountable to regulators or leadership. That changed how I approach every design decision, from information hierarchy to the smallest interaction state.

It also taught me what it means to design in a fast-moving product environment, where perfect research rarely exists and you have to make good calls quickly, with the signal you have. I got comfortable with that ambiguity, and better at moving fast without losing rigour.

All the work I did shipped to production across Conformity and Vendor Pulse, used by enterprise clients including HDFC Bank, Aditya Birla Capital, Deloitte, and Angel One. Both products are live and actively used across Zeron's enterprise client base. Designing for that calibre of enterprise clients early in my career shaped how I think about product quality.

Due to an NDA, detailed process documentation and all screens are not public, but the work is real and shipped.

GRC & VRM : Enterprise Product Design at Zeron

A graduation project designing a mobile-first smart home app that simplifies automation, surfaces real-time energy insights, and makes smart living effortless.

There is a paradox sitting at the centre of smart home adoption. People own the devices. They just do not really use them.

 of respondents in my survey owned at least one smart device. Only 

 had never once checked their energy usage.

That gap between ownership and meaningful use became the core design problem. The technology was not failing. The interfaces were.

To understand why that gap existed, I ran a 

, conducted 1:1 user interviews, and visited both smart home stores and actual homes where these systems were installed. Watching people interact with devices in real environments revealed something no survey response could: most users had settled into using 3 or 4 basic functions and quietly given up on everything else.

I also ran a competitive analysis across 

 Google Home, Amazon Alexa, Samsung SmartThings, Honeywell Home, Hubitat Elevation, and Home Assistant. Every one of them prioritised technical depth over behavioural clarity.

Three patterns came up consistently across all research:

 Automation setup was described as confusing and time-consuming. Most users had tried once, failed to understand the options, and never gone back.

 Without timely, contextual feedback, users had no real sense of how much energy they were consuming. A monthly bill was the only signal they received.

 Smart systems were not surfacing information when it was actually useful. Users had to know where to look and most did not.

 Highly tech-savvy, wants full automation and deep customisation. Frustrated by interoperability issues and features buried in menus. His bar for the product is high and nothing on the market is meeting it.

 Busy, family-oriented, wants the system to work without requiring ongoing configuration. Has given up on automation because the setup overwhelmed her. She does not need more features, she needs clarity.

 Lives alone and relies on smart home technology for independence and daily comfort. Struggles with small text, unreliable voice recognition in noisy environments, and interfaces that assume a level of technical confidence he does not have. Wants the system to work predictably without requiring him to troubleshoot or relearn it constantly.

These three users represented the full range of the problem. The solution needed to work for all of them without compromising for any of them.

Before moving into solutions, I framed the research into How Might We questions to bridge insight and ideation:

These became the filters for every design decision that followed.

Evio is a unified smart home control system built across five core experiences: a home dashboard, device management, energy analytics, routine automation, and a smartwatch companion for quick access.

The dashboard surfaces the most important information first: which devices are on, today's energy usage, the weekly sustainability score, and contextual quick actions. Everything actionable is reachable within two taps.

Devices are organised by room rather than by category, because that is how people mentally map their homes. Each device card shows real-time power consumption inline, making energy awareness a passive byproduct of normal use rather than something users have to seek out.

The analytics screen translates raw energy data into something meaningful. A weekly 

 compares your consumption against similar households and expresses the impact in human terms: this week your usage was equivalent to not running the AC for 2 hours. That framing makes the number feel real.

 surface peak usage patterns, missed automation opportunities, and efficiency tips based on actual device behaviour. Research showed AI-driven personalisation can reduce energy consumption by up to 

. The goal was to make that kind of intelligence feel accessible, not algorithmic.

The automation flow went through the most significant redesign of any feature in this project.

, 3 of them described it as too long with too many steps to get to one outcome. They completed the task but with visible hesitation and backtracking.

That feedback was specific enough to act on. I redesigned the flow to 

: select device, configure the automation on a single screen, review and save. Same functionality, substantially less cognitive load. All 3 users who had struggled with the original flow completed the new version without hesitation.

This was the clearest moment in the project where 

The smartwatch companion was scoped deliberately to essentials: active devices, room-based navigation, and quick controls for the most common functions. Every interaction on a small screen has to justify its presence.

Rather than mirroring the full mobile experience, the watch interface was designed to be glanceable and purposeful. 

 completed the watch task successfully on the first attempt and described it as straightforward and easy to navigate.

10 users tested across varied ages and tech familiarity levels across 4 tasks:

 3 users found the original flow too long. Directly informed the Phase 6 redesign.

 No friction. Users appreciated the clarity of role permissions.

 Some users explored the interface before acting. Largely positive.

 Described as intuitive. Limiting scope to essentials worked.

The hardest design problems in consumer technology are not visual ones. They are behavioural.

The people I researched were not disengaged. They wanted to use their smart homes more fully. The gap between intention and action existed almost entirely because of design friction: too many steps, too little feedback, too much assumed knowledge.

Evio was an attempt to close that gap. Not by adding features but by removing the things that made existing ones feel unreachable.

I also learned something about testing that I will carry into every future project: it is a discovery tool, not a checkpoint. The routine redesign did not come from a heuristic review or a design critique. It came from watching three real people hesitate at exactly the same moment. That is information no wireframe can surface.

Explore the full graduation project book:

https://drive.google.com/file/d/1RdM0fAN84JAvw7WwA0O48D_MjNmKIybL/view?usp=drive_link

Evio: Designing a Smart Home Interface for Better Living and Energy Efficiency

A UX audit of BigBasket.com using Nielsen Norman Group's 10 heuristics and Think Aloud user testing to uncover friction in grocery e-commerce navigation.

Every audit needs a specific task to anchor it. I chose: 

navigate the BigBasket website to find a pain reliever and add it to cart

. A routine, everyday task. It turned out to be significantly harder than it should have been.

Nielsen Norman Group's 10 usability heuristics

, the industry standard framework for evaluating interface usability. To reduce individual bias, I recruited 

 who evaluated the website independently alongside me. Each evaluator worked through the heuristics separately before we compared findings, 

 overlapping issues across heuristics and evaluators.

 A real user completed the same task while verbalising their thoughts throughout, followed by a structured 

 to understand their reasoning and pain points. The goal was to validate what the heuristic evaluation had flagged and surface anything the expert review had missed.

The evaluation surfaced issues across all 

 Pharmaceutical products were housed under 

, a label no user would associate with medicines. This single naming decision broke the most common pharmacy-related task on the platform.

 Buttons, icons, and styles varied across pages with no clear system. Some buttons were rounded, some were not. Icons used different visual languages in different sections.

 Reaching help and support required multiple clicks with no clear path. Users who ran into problems had no quick way out.

 Once inside a subcategory, there was no clear way to step back to the parent category. Users had to start over entirely.

 CTAs were placed in unexpected positions. On the checkout page, the primary action sat above the cart items rather than below them, working against natural reading flow.

 session involved one participant completing the task while 

That last quote says everything. A user on one of India's largest grocery platforms began questioning whether the platform sold medicines at all, because of one mislabeled category. 

When the evaluation and real user behaviour point to the same problems, the case for fixing them becomes very hard to ignore. Both methods converged on: 

confusing category nomenclature, poor information hierarchy, missing backtracking options, and inconsistent UI elements.

 Pharmaceutical products need a clearly labelled dedicated section. "Beauty and Hygiene" should not be the home for medicines.

 Users need a consistent way to move between subcategories and parent categories without starting from scratch.

 Buttons, icons, and interactive elements need consistent styling across the entire platform.

 CTAs belong where users expect them. Critical actions should be impossible to miss.

This project made clear how much a single naming decision can affect an entire task flow. One mislabeled category caused a user to lose their way completely and start doubting the platform itself.

Pairing heuristic evaluation with user testing also changed how I think about research. The expert review identified what was wrong. The think aloud session showed how it actually felt to experience that wrongness. Seeing both converge on the same issues made the findings far more credible than either method would have on its own.

 This was a research and audit project. The output was a set of recommendations, not redesigned screens. The value here is in the methodology and the depth of the findings.

BigBasket.com: Heuristics and Usability Testing

A B2B SaaS product design project building a multi-role skill management system for SMEs, bridging skill gaps and reducing employee turnover through upskilling.

, yet most do not have access to talent management tools that actually fit how they operate. Enterprise HR platforms are too expensive, too complex, and built for organisations five times their size. This project was about designing something that worked for the middle ground: medium-sized businesses with 100-500 employees, dealing with real skill gaps, high turnover, and no great options.

Mid-sized businesses are caught in a difficult position when it comes to talent. They are too large for informal, relationship-based management and too resource-constrained for the tools built for enterprise. The result: skill gaps widen, employees feel undervalued, and turnover climbs.

The problem statement that came out of the research:

"Mid and Small sized businesses face a complex talent management maze. Manual processes and outdated tools hinder upskilling and smooth hiring. This creates skill gaps, slows workflows, and drives up recruitment costs. Employees crave growth opportunities. Feeling undervalued and overloaded, they leave."

 across HR Managers and founders of MNCs and startups. The interviews were more revealing than the surveys as they gave honest, specific answers about what was actually failing inside these organisations.

 of companies surveyed were facing active talent retention problems.

 identified upskilling as a significant challenge.

 of employees were experiencing monotony in their roles.

 of the organisations were still managing talent processes manually.

Three quotes that shaped the direction of the entire product:

The last quote pointed to something important: the problem was not just skill gaps. It was visibility. Employees and managers both lacked a clear picture of where skills stood and where they needed to go.

The research shaped three distinct archetypes:

 Needs data-driven workforce insights, efficient tools for hiring and development, and platforms that employees will actually adopt. Frustrated by low software engagement across teams.

 Needs skill-based task allocation, team development tracking, and a way to raise training needs based on what they observe day to day. Balancing project deadlines against team growth.

 Needs visibility into their own skill gaps, clear paths for upskilling, and recognition for learning efforts. Wants to grow but often lacks the tools or encouragement to act on it.

Synergy is a B2B SaaS skill management platform built around a skills-first approach. Rather than treating training as a periodic HR activity, Synergy makes skill development a continuous, visible, and rewarding part of daily work.

, each serving a distinct user without overwhelming them with information meant for someone else.

Employees can view their current skills against expected levels through a skill gap analysis, see personalised training suggestions based on their profile, raise training requests, track their progress, and earn points through a team leaderboard. The gamification was deliberate: it made upskilling feel like participation rather than obligation.

The platform also included a mobile app designed specifically for employees. The onboarding flow collected interest areas and career satisfaction levels upfront through a 4-step personalised setup, so training recommendations felt relevant from the first login rather than generic.

On the dashboard, employees could check their leaderboard position, track points, manage current tasks, and redeem rewards directly from different brands. Making the rewards tangible and immediately accessible was a deliberate choice to keep engagement with upskilling feeling worthwhile outside of office hours.

Managers can view their team's skill profiles, identify gaps, raise training requests, and forward feedback to HR or the training department. The HR dashboard provides a company-wide view of training activity, employee details, and statuses. The training dashboard allows the training team to schedule, approve, and track trainings across teams.

Building four role-specific interfaces rather than one universal dashboard was the most important structural decision in the project. Each user sees only what is relevant to their responsibilities.

One of the most meaningful changes in this project was a vocabulary decision. Replacing the word "metrics" with "milestones" across the employee dashboard completely shifted how the data felt. Metrics implied being measured. Milestones implied progress being celebrated. Same data, different emotional relationship with it.

It reinforced something about B2B product design: how you name things shapes how people feel about using them.

Designing for SMEs meant sitting with a tension that does not resolve cleanly: these organisations need structure, but too much structure kills the flexibility they rely on. Every feature had to earn its place by being genuinely useful rather than just comprehensive.

B2B tools carry more emotional friction than they are given credit for. People use them daily, and when an interface feels intimidating or cold, it directly affects how much they engage with it. Role-specific design is one way to reduce that friction. Giving people only what is relevant to them builds trust faster than giving them everything at once.

Synergy: Designing a Skill Management Platform for SMEs

It has been great to see the creativity and attention to detail by Navya. Her design skills and dedication have been remarkable. Joining as a UI/UX intern, within a short period, she was able to work with the team efficiently and provide creative suggestions for our SaaS products. I am sure that her great attitude and eye for perfection will take her places.

Bhabani Panda

Navya is a diligent candidate with exceptional knowledge in the UX and Design domain. Her ability to bring out her creativity in her work output is remarkable and she has a great attitude towards her work.

Murali Sunderraghavan

I’ve had the opportunity to manage Navya as a UI UX Design Associate at Zeron, and one thing that consistently stood out was her ability to break down complex problem statements and translate them into intuitive, user focused design solutions.

She has a strong understanding of UX thinking, pays close attention to detail, and brings a thoughtful design approach even while working on fast moving product requirements. Along with her design craftsmanship, she’s proactive, collaborative, and quick to adapt, which made her a dependable part of the team.

Navya has solid potential as a product designer, and I’d confidently recommend her to any team looking for someone who combines good design sense with problem solving ability and a strong learning mindset.

Shaikh Dastagir

Navya is an exceptional UX Engineer with the ability of turning ideas into impactful realities. Our collaborative brainstorming sessions are not only productive but also enjoyable, thanks to her contributions. I affectionately refer to Navya as an "annoyingly picky" designer, a term that truly reflects her dedication.Throughout our projects together, Navya has consistently enhanced the quality of our work. Working with her on The Walls was by far the most fun and creative space I've been in, The Walls did not have any muscle, any clothes but only bones until Navya joined. And she's just fun to be around.

Indigo VUI: Designing a Voice Interface for Flight Booking

Navya Sharma