Crafting intuitive, scalable experiences ✨ through AI‑native workflows driving efficiency and outcome‑oriented design.
A User Experience Designer with 15+ years of experience across SaaS, enterprise, and AI-driven products working across research, strategy, and design to solve complex problems in a simple, human way. I also bring AI into my UX workflows using tools like Figma Make, ChatGPT, Claude, and Gemini to speed up ideation, streamline processes, and make better design decisions.
Recent Works
ZZAZZ - Content Economy, Redesigned
Designing a dynamic pricing ecosystem for the open content economy - Terminal, Publish, and Signal powered by LPM.
UX ResearchProduct DesignDesign SystemPrototyping
SalesX - Voice-First CRM
Designing a simplified AI/ML-based CRM that eliminates data entry friction for modern sales teams.
UX ResearchVisual DesignDesign SystemPrototyping
Airbus - Aircraft Ground Operations
Mission-critical interface design for A350 and A380 aircraft maintenance and authorisation systems.
ZZAZZ is building the infrastructure layer for the content economy - a system that replaces today’s broken creator monetisation model with one that’s dynamic, fair, and friction-free on both sides. Three products form the ecosystem: Terminal (consumer discovery and access), Publish (creator publishing and analytics), and Signal (a lightweight embeddable widget). Powering all three is LPM - a Large Pricing Model that sets real-time access prices based on demand, engagement, and context.
My role was to translate this ambitious technical vision into product experiences that felt natural to use, from the first search query to the first payout.
01Context & Challenge
Team & Tools
Team
3
Product Managers
5
Product Designers
Including me, across all three products.
Gemini
3 days
Compressed a 2-week research phase via competitor teardowns and paywall UX sweeps.
Claude + MCP
<4 hours
PRD drafts, user stories, and API specs - plus Jira and Confluence auto-populated each sprint.
Figma Make
15 min
Signal widget variants per iteration round - down from ~2 hrs of manual design.
Full Cycle
4 weeks
Scoped as 10 weeks. No cuts to quality.
The Problem
Creators are locked into subscription-or-free binaries that leave money on the table, while readers drown in recurring bills just to follow the writers they like. Neither side wins - the market needed a new pricing primitive.
Where the system breaks - creators can't price their work fairly, and readers won't pay through friction-heavy subscription walls.
Creator side
All-or-nothing monetisation
Poor trial-to-paid conversion
No visibility into what drives willingness to pay
Platform dependency and algorithm risk
Consumer side
Subscription fatigue from recurring charges
No way to pay for a single article
Inconsistent access across publishers
No unified reading identity or history
The ZZAZZ Approach
ZZAZZ introduces the Large Pricing Model (LPM) - a dynamic pricing layer that replaces the binary free/paid choice with a full pricing spectrum set in real time, based on content demand, reader engagement history, and context signals. Readers pay only for what they actually want to read; creators get a tuned, automated monetisation layer without configuration overhead. Three interconnected products make this work in practice:
Product 01
Terminal
Consumer-facing search and discovery. Readers find content across all ZZAZZ publishers, see real-time pricing, and pay per piece - no subscription required.
Product 02
Publish
Creator-facing publishing dashboard. Writers manage their catalogue, configure pricing ranges, monitor earnings, and view audience analytics in one place.
Product 03
Signal
An embeddable website widget that publishers drop onto existing sites. Signal surfaces real-time pricing and enables one-click access without redirecting readers away.
My design scope
Across all three products, my primary responsibility was the reader-facing experience: Terminal's discovery, pricing transparency, and first-payment flow. I co-designed Publish's creator workspace and owned the Signal widget interaction pattern end-to-end.
What I was asked to solve
The founding team had a strong technical thesis (LPM as a pricing primitive), but no answer to the design question: how do you make dynamic pricing feel fair rather than arbitrary to someone who has never encountered it before? That was the brief I started with.
02Research
Research Plan
01Stakeholder InterviewsBusiness goals, constraints, and LPM value proposition with founders and product leads.
02User InterviewsDepth sessions with independent writers and frequent readers on paywalls and payment friction.
03Competitive AnalysisAudit of Substack, Medium, Patreon, YouTube, and news paywalls to map the monetisation landscape.
No major platform combines creator pricing control with flexible per-access reading. Substack, Medium, Patreon - all binary: free or subscription.
The Opportunity
High creator control + flexible reader access is entirely unoccupied. That quadrant is ZZAZZ’s core positioning.
The Differentiator
Dynamic pricing (LPM) + embeddable widget (Signal) + cross-publisher reader identity - none of the five competitors offer all three.
Survey - 142 responses across creators and readers, validating qualitative findings at scale.
142
respondents
76%
hold 3+ subs, actively read fewer than 2
83%
abandoned a paywall in the past two weeks
61%
of creators earn under ₹10k/month despite 15+ hrs/week
54%
would pay ₹5–20/article with no sign-up, under 10s
71%
open to AI pricing if they keep a floor price override
89%
want to see why a price was set, not just what it is
Four research tracks converged on the same gap: no platform gave creators pricing control and readers a frictionless way to pay per article simultaneously. That gap wasn't a feature request, it was the entire product opportunity. Everything that followed was designed to close it.
The most important pattern - Identity & Continuity - wasn't in my original screener at all. It changed the scope.
03Analysis & Convergence
Key Insights
01
Subscription fatigue is about commitment, not cost.
02
Creators don’t want more configuration - they want better defaults.
03
Dynamic pricing is only trustworthy if readers can see why the price was set.
04
Signal must feel native to the publisher, not ZZAZZ.
05
Cross-publisher reading history - the feature readers didn’t know they wanted.
Design Direction
Decision 01
Transparency before features
89% of readers wanted to know why a price was set. Pricing legibility became a structural requirement across all three surfaces from day one.
Decision 02
Terminal as the north star
The reader’s first-payment moment was the stress-test for every systemic decision. Publish and Signal were designed in parallel, but Terminal came first.
Decision 03
Friction reduction over configuration
83% of paywall abandonment was due to steps, not price. Any feature adding overhead before a first payment or payout was deprioritised.
04Synthesis & Divergence
Personas
Three distinct users, one shared constraint: trust.
AS
Arjun Sharma
Independent Writer · Publish
Goal
Sustainable income from writing. Needs a middle ground between free and hard paywall.
Pain
Platform fees eat 16% before payout. No pricing flexibility between free and paid.
₹2,400avg. monthly - 64% below viability
DC
Divya Chinnappa
Product Manager · Terminal
Goal
Quality content on demand. Support the writers she values without another monthly commitment.
Pain
Pays for 4 subscriptions, actively reads 1. Every article hits a paywall anyway.
73%of paid subs go unread each month
RP
Rajesh Pillai
Digital Editor · Signal + Publish
Goal
Loyal readership beyond algorithmic reach. Flexible monetisation without hard paywalls.
Pain
AI answers queries before readers reach the site. Hard paywalls kill discovery.
−41%organic search traffic in 18 months
Journey Maps
Following Arjun across six stages - from discovery to the loyalty flywheel.
The design-critical path through Terminal - the reader’s first-payment moment.
How Might We
HMW 01
Make dynamic pricing feel fair to readers who’ve never encountered it before?
HMW 02
Let creators benefit from LPM without adding configuration to their workload?
HMW 03
Design Signal to feel native to the publisher’s brand, not an overlay?
HMW 04
Build a cross-publisher reading identity that feels like a benefit, not tracking?
HMW 05
Make the first payment fast enough that readers don’t abandon at the decision point?
Three personas, three surfaces, but one shared constraint: trust. The priority matrix forced a clear order. First-payment flow and pricing legibility ranked above every configuration feature. Anything that added steps before a reader paid or before a creator understood their earnings was deprioritised, regardless of how requested it was.
05Design & Iteration
Design Principles
Principle 01
Legible pricing
Every price must carry enough context to explain itself - no hunting required.
89% of readers wanted to know why a price was set. 6/8 missed it entirely when unstyled.
Principle 02
Zero friction at payment
Auth and payment front-loaded on first use. Every access after that is one tap.
83% abandoned a paywall recently. The barrier wasn’t price - it was the number of steps.
Principle 03
Creator confidence, no config overhead
LPM runs in the background. Creators set a floor and ceiling - that’s it.
4/5 creators weren’t opposed to AI pricing - they just needed an override. Guardrails, not a panel.
Design Decision: QAP Chart
A 72-hour price history chart - like a stock sparkline - shown at the paywall. Readers could see if the price was rising, falling, or stable. Context made dynamic pricing feel fair rather than arbitrary.
What it changed
In V1, 5/8 readers called dynamic pricing “random.” After the QAP chart in V2, that dropped to 1. It didn’t add trust - it made the existing trust logic visible.
The pushback
“Readers aren’t traders.” My counter: you don’t need financial literacy to understand a falling line means the price is dropping. Familiarity is the mechanism.
Design System
Designing across three products simultaneously made one thing clear: a shared system wasn't optional. Every component had to carry meaning in three different contexts: discovery, publishing, and embedding. The design language that emerged wasn't a style guide; it was the connective tissue that made Terminal, Publish, and Signal feel like one ecosystem rather than three separate tools.
06Testing & Validation
Usability Testing
Method
Moderated remote
Think-aloud · Loom · Google Meet
Participants · n=8
4Independent creators
4Paid newsletter readers
Task Flows · 3
T1Find & buy a paywalled article
T2Set paywall break in Publish
T3Subscribe via Signal mid-article
Success Metrics
Completion rate≥80%
SUS score≥75
Time on taskvs. baseline
Error frequencyper path
87%
Avg. task completion
78
SUS score (Good)
−34%
Time on task vs. V1
3
Critical issues found
T1 · Price DiscoveryCritical
Finding · Root cause
6/8 missed the price tag - price badge was small grey text inside meta row, same weight as author name.
Fix applied
Promoted to coloured pill badge on article card. Completion rate 50% → 87%.
T2 · Paywall BreakModerate
Finding · Root cause
3/4 creators couldn't find the paywall break control - buried 2 levels deep in "Monetise" dropdown, icon-only.
Fix applied
Inline block-level insertion via + between paragraphs. Discovery time 94s → 18s.
T3 · Signal TrustModerate
Finding · Root cause
2/4 readers described the widget as "an ad" - generic "Subscribe" CTA with no creator context in the collapsed pill.
Three critical issues. All three fixed before handoff. The clearest signal from testing: pricing transparency wasn't a trust-builder, it was a prerequisite. Without it, readers stalled before they even reached a payment decision. Small copy and hierarchy changes moved completion rates by margins that no visual redesign alone would have achieved.
BreakingFed Chair Powell Signals Rate Cuts Could Come Sooner Than Expected, Sending Markets to Record HighsUpdated 2m ago
Analysis
The A.I. Race and the Specter of the Dot-Com Boom: ‘We’ve Heard This Story Before’
Investors are pouring billions into artificial intelligence. Historians and economists are asking whether the euphoria will end differently than the last great technology bubble.
4 Min Read
$ 4.89
Top 1% in Politics
$ 4.89USD
Live Price
AdTimepay• 5s
Live pricing based on demand. Learn more
A.I. Regulation Bill Advances After Rare Bipartisan Senate Vote
6 Min Read
Big Tech Orders Full-Time Office Return as Remote Work Era Fades
5 Min Read
Opinion
The Attention Economy Is Quietly Reshaping How We Think About Power
By Ezra Klein
7 Min Read
Tech Giants Post Record Earnings as A.I. Revenue Surges Past Forecasts
3 Min Read
$ 3.45
Top 5% in Technology
$ 3.45USD
Live Price
AdTimepay• 5s
Live pricing based on demand. Learn more
Supreme Court to Hear First Major Case on A.I.-Generated Art and Copyright
4 Min Read
$ 2.87
Trending in Law & Policy
$ 2.87USD
Live Price
AdTimepay• 5s
Live pricing based on demand. Learn more
Traders on the floor of the New York Stock Exchange as the Dow surged 800 points Thursday.
Richard Drew / AP Photo
New York, N.Y.
LIVE5m ago
Dow Surges 800 Points as Fed Chair Signals Rate Cuts Could Come Sooner Than Expected
Federal Reserve Chair Jerome Powell struck a more dovish tone, sending stocks to record highs and sparking a broad rally as investors priced in eased borrowing costs ahead.
See more updates ›
Taylor Swift’s Eras Tour Becomes the Highest-Grossing Concert Tour in History
5 Min Read
OpenAI Unveils New Model It Says Can Reason Like a Scientist, Stirring Debate
3 Min Read
Wikimedia Commons
Alex Soros and Huma Abedin Share Their Love Story and Wedding
Each knew of the other, but it took a conversation at a party - plus a missed DM, many texts, calls and the Doha Forum - to set a dinner date.
8 Min Read
Dallas Cowboys Cheerleaders Get a 400 Percent Pay Raise
5 Min Read
Why Is Everybody ‘Crashing Out’? The Rise of a Useful Slang Term.
3 Min Read
Fed Holds Rates Steady, Signals Two Cuts Are Still Possible in 2025
4 Min Read
$ 0.99
Top 10% in Finance
$ 0.99USD
Live Price
AdTimepay• 5s
Live pricing based on demand. Learn more
Netflix Adds 8 Million Subscribers as Password Crackdown Continues to Pay Off
3 Min Read
$ 1.31
Top 5% in Entertainment
$ 1.31USD
Live Price
AdTimepay• 5s
Live pricing based on demand. Learn more
Trump AdministrationDomestic Policy BillMonetized PresidencyNew Political LandscapeApproval Ratings
U.S.INTERNATIONALCANADA
The New York Times
U.S. ▾World ▾Business ▾Arts ▾
ANALYSIS
The A.I. Race and the Specter of the Dot-Com Boom
Investors are pouring billions into artificial intelligence. Historians are asking whether the euphoria will end differently.
$ 4.89
Tech Giants Post Record Earnings as A.I. Revenue Surges
3 MIN READ
Fed Holds Rates Steady, Signals Two Cuts Possible in 2025
4 MIN READ
$ 4.89
Live Price
Top 1% in Politics
Timepay byZZAZZ
Technology Analysis|
LIVEMarkets2m ago
|
A.I. RegulationFed PolicyTech Earnings
Technology Analysis
The A.I. Race and the Specter of the Dot-Com Boom: ‘We’ve Heard This Story Before’
Investors are pouring billions into artificial intelligence. Historians and economists are asking whether the euphoria will end differently than the last great technology bubble.
By David StreitfeldJune 19, 2025
4 MIN READ
Listen · 6:42 min
Share full article
412
Traders on the floor of the New York Stock Exchange, where A.I.-linked stocks have surged more than 300 percent since 2023. Spencer Platt/Getty Images
SAN FRANCISCO - In 1999, a company called Pets.com raised $82.5 million in an initial public offering, burned through most of it in nine months and collapsed in what became one of the most vivid symbols of the dot-com bust. Today, venture capitalists are betting similar sums on A.I. startups that have yet to turn a profit - and some on Wall Street are beginning to ask uncomfortable questions.
The numbers are staggering. More than $200 billion flowed into artificial intelligence investments last year alone, according to data from PitchBook, dwarfing the peak capital deployed during the late 1990s technology boom even after adjusting for inflation. Nvidia, whose chips power most large language models, briefly became the world’s most valuable company, with a market capitalization exceeding $3 trillion.
The comparison to the dot-com era is not universally accepted. Proponents of A.I. argue that today’s companies have genuine revenue, real customers and transformative technology - unlike the speculative ventures of 1999. OpenAI is on track to generate more than $4 billion in annualized revenue. Microsoft has embedded A.I. across its product suite. Google’s Gemini is used by hundreds of millions of people.
But skeptics point to patterns that feel familiar. Valuations have become untethered from earnings. Executives at even pre-revenue startups speak of “inevitable” dominance. And the word “transformative” - a staple of 1990s prospectuses - appears in nearly every A.I. pitch deck.
“Every bubble has a kernel of truth,” said Bradford DeLong, an economic historian at the University of California, Berkeley. “The internet really did transform the world. But that didn’t mean Pets.com was worth anything. The same logic applies here. The technology may be real. The valuations may still be fantasy.”
What happens next, economists say, depends partly on whether A.I. can actually deliver the productivity gains that justify the investment. Early signs are mixed. Some industries - law, software development, customer service - are already seeing measurable efficiency gains. Others have found that the technology introduces as many complications as it solves.
“Every bubble has a kernel of truth. The technology may be real. The valuations may still be fantasy.”
Tech Giants Post Record Earnings as A.I. Revenue Surges Past Wall Street Forecasts
Microsoft, Alphabet and Meta reported their strongest quarters in years, driven almost entirely by artificial intelligence products and cloud infrastructure spending.
By Erin GriffithJune 19, 2025
3 MIN READ
Listen · 4:18 min
Share full article
289
Microsoft’s Azure A.I. division accounted for nearly 40 percent of the company’s cloud revenue in the first quarter of 2025. Microsoft Corp.
SAN FRANCISCO - The five largest technology companies in the United States posted a combined $95 billion in quarterly profit on Tuesday, shattering analyst expectations and confirming that artificial intelligence has become the defining growth engine of the global economy - at least for now.
Microsoft reported revenue of $71.2 billion, up 22 percent from a year earlier, fueled almost entirely by its Azure cloud platform and its Copilot suite of A.I. tools now embedded in Office, Teams and Windows. The company added that it would spend $80 billion on data center infrastructure this year, nearly double its capital expenditure from 2023.
Alphabet, Google’s parent company, posted earnings of $26.3 billion, up 31 percent, and raised its full-year revenue guidance for the third consecutive quarter. Chief Executive Sundar Pichai told analysts that A.I. Overviews - the A.I.-generated summaries that now appear at the top of Google Search - are used by more than 1.5 billion people each month.
Meta’s results were perhaps the most striking. The company behind Facebook and Instagram reported $17.8 billion in net income, a 47 percent increase, driven by an A.I.-powered advertising system that the company says can now predict with remarkable accuracy which ads users are most likely to click. Chief Executive Mark Zuckerberg called it “the most profitable year in Meta’s history, with more to come.”
The results sent tech stocks surging after hours, with the Nasdaq Composite briefly touching a record high. But some analysts cautioned that the earnings masked a dependence on capital spending that could become problematic if A.I. revenue growth slows. Together, the five tech giants spent more than $200 billion on A.I. infrastructure last year - money that will need to be justified by commensurate returns.
“The results confirmed that A.I. has become the defining growth engine of the global economy - at least for now.”
Supreme Court Agrees to Hear First Major Case on A.I.-Generated Art and Copyright
The case, which involves an image-generation company and a coalition of artists, could determine whether works created by artificial intelligence are entitled to copyright protection.
By Adam LiptakJune 19, 2025
4 MIN READ
Listen · 5:33 min
Share full article
531
The Supreme Court agreed to hear a case that could reshape the future of copyright law in the age of artificial intelligence. AP Photo/J. Scott Applewhite
WASHINGTON - The Supreme Court agreed on Thursday to hear its first major case involving artificial intelligence and copyright, a dispute that could reshape creative industries and determine who - if anyone - owns the output of image-generation software.
The case, Midjourney v. Artists’ Coalition, centers on whether A.I. companies that train their systems on billions of copyrighted images without permission have engaged in infringement, and whether the images those systems produce are themselves eligible for copyright protection. Lower courts have issued contradictory rulings, creating a patchwork of legal uncertainty.
A coalition of visual artists sued Midjourney, Stability AI and several other companies in 2023, arguing that their work had been used without consent or compensation to train models that now compete directly with human artists. The companies have argued that training on publicly available images constitutes fair use - a legal doctrine that permits certain uses of copyrighted material without permission.
The copyright question is equally thorny. The Copyright Office has maintained that works created solely by A.I. - without sufficient human creative input - are not eligible for copyright protection. But what constitutes “sufficient” human creativity when a person types a prompt that generates an image is deeply contested.
“This case will define the creative economy for the next century,” said Rebecca Tushnet, a copyright law professor at Harvard Law School. “The court has to answer questions that Congress never anticipated when it drafted copyright law, in a world where machines can generate thousands of images a second.”
“This case will define the creative economy for the next century. The court has to answer questions Congress never anticipated.”
Fed Holds Rates Steady, Signals Two Cuts Still Possible in 2025 Despite Inflation Uncertainty
Federal Reserve officials left interest rates unchanged for the fourth consecutive meeting, but their updated projections suggested a narrow path to easing before year’s end.
By Jeanna SmialekJune 19, 2025
4 MIN READ
Listen · 5:07 min
Share full article
198
The Marriner S. Eccles Federal Reserve building in Washington. The Fed left its benchmark interest rate unchanged at its June meeting. Federal Reserve/Public Domain
WASHINGTON - Federal Reserve officials voted unanimously on Wednesday to hold their benchmark interest rate steady at 4.25 to 4.5 percent, citing continued uncertainty about the inflation outlook while leaving the door open to rate cuts later in the year if economic conditions allow.
The decision was widely expected after a series of economic reports showing inflation cooling gradually but unevenly. The consumer price index rose 2.6 percent in May compared with a year earlier, down from a peak of 9.1 percent in 2022 but still above the Fed’s 2 percent target. Core inflation, which strips out volatile food and energy prices, remained at 2.8 percent.
In their updated economic projections, known as the Summary of Economic Projections or “dot plot,” Fed officials indicated they expected to cut rates twice before the end of 2025, though with considerably less certainty than they expressed six months ago. The median projection showed the federal funds rate ending the year at 3.75 to 4 percent.
Jerome H. Powell, the Fed chairman, was characteristically cautious at his post-meeting news conference. “We are making good progress,” he said, “but we are not yet at a point where we are confident that inflation is durably on a path back to 2 percent. The economy remains strong. Labor markets remain solid. We have the luxury of time.”
Markets responded with a modest rally. The S&P 500 rose 0.6 percent after the decision, while yields on two-year Treasury notes, which are highly sensitive to Fed policy, fell slightly. Traders in federal funds futures markets slightly increased their bets on two rate cuts this year, pricing in roughly a 60 percent chance of the first cut arriving in September.
“We are making good progress, but we are not yet confident that inflation is durably on a path back to 2 percent.”
Federal ReserveInterest RatesInflationEconomyJerome Powell
Unlocked via TimePay · $0.99
Entertainment|
LIVEMarkets2m ago
|
A.I. RegulationFed PolicyTech Earnings
Entertainment
Netflix Adds 8 Million Subscribers as Password Crackdown Continues to Reshape Streaming
The company’s aggressive paid-sharing initiative, once feared to anger users, has instead driven its strongest subscriber growth in three years.
By Brooks BarnesJune 19, 2025
3 MIN READ
Listen · 3:55 min
Share full article
347
Netflix reported its best quarterly subscriber growth since 2021, fueled by its paid password-sharing program and a strong content slate. Netflix/Wikimedia Commons
LOS ANGELES - Netflix reported Tuesday that it added 8.05 million subscribers in the second quarter, far exceeding analyst forecasts and cementing the company’s position as the dominant force in a global streaming market that rivals have struggled to monetize.
The results surprised Wall Street, which had expected Netflix to add roughly 5.5 million subscribers after a strong first quarter. Shares of Netflix jumped more than 11 percent in after-hours trading, adding roughly $25 billion to the company’s market capitalization.
The growth was driven in large part by Netflix’s paid password-sharing initiative, which the company began rolling out globally in mid-2023. The program, which requires users to pay an additional fee of $7.99 a month to share their account outside their household, initially sparked a backlash. But the feared mass cancellations never materialized, and instead, millions of former freeloaders converted to paying customers.
Netflix’s advertising-supported tier, which costs $6.99 a month, also gained significant traction. The company said the ad-supported plan now accounts for 45 percent of all new sign-ups in markets where it is available, up from 30 percent a year ago. Netflix said it expected advertising revenue to double year-over-year in 2025.
Greg Peters, Netflix’s co-chief executive, credited the company’s content slate for sustaining momentum. “When you have great shows, people sign up and they stay,” Mr. Peters said on a call with analysts. “The content machine is firing on all cylinders.” Recent hits have included a new season of “Stranger Things,” a documentary on the 2024 Olympic Games and a live comedy special that drew 28 million concurrent viewers.
“When you have great shows, people sign up and they stay. The content machine is firing on all cylinders.”
Across 15 months and three shipped products, the design work moved three numbers that mattered, and surfaced one that I hadn't anticipated.
87%
Task completion
Across all three usability test flows. Up from 51% average in V1, driven almost entirely by the price badge redesign and paywall break discovery fix.
−34%
Time on task vs. V1
The paywall break control went from 94s to 18s discovery time after I moved it inline. No feature change, just a placement decision.
78
SUS score (Good)
Industry average for SaaS is 68. At 78, Terminal and Publish cleared "Good" territory on first round of testing, before further iteration.
3
Pilot publishers
Signal was adopted by three pilot publishers within the first month after the trust redesign. Adding the creator name to the collapsed widget pill was the fix that unblocked them.
The number I didn't anticipate: of the three critical issues found in testing, all three were copy and hierarchy changes, not component rebuilds, not new flows. The biggest usability gains came from making existing information more visible, not from adding new information. That shaped how I approach every audit I run now.
What I'd Do Differently
Learning 01
Explain novelty inline, not upfront.
Onboarding screens had the worst recall - only 2/8 remembered the pricing explanation 15 mins later. Inline copy at the moment of encounter had 4× better recall.
Learning 02
Same tokens, different affordances per product.
Identical article cards across Terminal and Publish confused readers - the card read as “your content,” not “content to discover.” Shared tokens, distinct interaction patterns per surface.
Learning 03
Design the entry point last, not first.
Signal was designed last but was the first touchpoint for most readers. 3/4 who hit it first described ZZAZZ as “a widget thing.” Sequence effort by user-facing impact, not engineering complexity.
Redesigning CRM from the ground up - powered by NLP (Natural Language Processing) to eliminate manual data entry through voice-first interaction so sales teams can focus on what they actually do best.
ClientSalesX (salesx.io)
Duration2 years · 2017–2019
RoleLead UX Designer
ScopeUX Research, Wireframing, Prototyping, UI Design, Design System
TechnologyNLP (Natural Language Processing)
SalesX is reinventing the CRM experience through NLP-powered voice interaction - allowing sales teams to capture, log, and manage data entirely by voice, and focus purely on selling. Traditional CRMs have evolved into systems that demand excessive manual data entry, often shifting the focus of relationship managers from client engagement to administrative tasks.
The goal: create a self-driving CRM that captures, organises, and analyses sales activities automatically, making manual data entry obsolete.
Zero Data Entry & Automated Organisation
Intelligent Meeting Summaries & Smart Follow-ups
Effortless Pipeline Management
01
Understanding the user
User research · Personas · Problem statements
User Research
What we learned in the field.
Our user research revealed that sales professionals struggle with excessive manual data entry, context switching across multiple tools, and inefficient follow-up management. SalesX addresses these pain points by automating data capture, organising interactions intelligently, and providing AI-driven insights to streamline the sales workflow.
8In-depth interviews with sales professionals across different industries
3Companies observed during a 1-week sales team shadowing study
50+Sales professionals surveyed about their daily workflow challenges
1Deep analysis of existing CRM usage patterns and failure modes
Pain Points
Three critical breakdowns in how CRMs are used today.
01
Manual Data Entry Overload
Sales professionals spend excessive time logging interactions, with 82% missing crucial details and 67% delaying note entry, leading to irretrievable information loss.
82%miss crucial details
67%delay note entry
02
Context Switching Disruptions
Managing customer relationships requires juggling multiple tools, causing 73% to lose focus mid-task and 89% to desire automatic activity logging.
73%lose focus mid-task
89%want auto-logging
03
Inefficient Follow-up Management
Manual tracking leads to 64% missing deadlines, 91% wanting AI-driven reminders, and 78% struggling to maintain accurate customer histories.
64%miss deadlines
91%want AI reminders
Personas
Who we designed for.
Enterprise Account Executive
Sarah Chen
Age 35 · San Francisco
Sarah manages large enterprise accounts at a B2B SaaS company, handling deals worth $500K+. Tech-savvy but values efficiency over feature complexity.
Goals
Minimise time on administrative tasks
Track complex stakeholder relationships
Maintain detailed meeting records without manual effort
Frustrations
"I spend more time updating the CRM than talking to customers"
"Important details get lost between meetings"
"Can’t quickly find historical context during calls"
40%
of selling time lost to CRM admin - blocking her $1.5M monthly quota
Sales Manager
David Rodriguez
Age 42 · Chicago
David leads a team of 12 sales reps, focusing on team performance, forecasting, and process optimisation. Data accuracy is his greatest daily blocker.
Goals
Accurate pipeline visibility
Team productivity optimisation
Data-driven coaching for reps
Frustrations
"Can’t trust the data in our CRM"
"Reps don’t update their opportunities regularly"
"Difficult to identify coaching opportunities"
60%
of CRM data is outdated - forcing 15+ hrs/week on validation instead of coaching
System works in the background without user intervention
Information appears automatically when needed
Zero-click information capture
Contextual Intelligence
Smart categorisation of interactions
Automated relationship insights
Predictive next actions
Minimal Interaction Required
One-tap access to key functions
Automated data organisation
Smart defaults based on user behaviour
References
Products that shaped our thinking.
Superhuman & Apple Live Activities
Auto-prioritisation of important emails and seamless interactions. Subtle background updates without user intervention - context-aware recommendations that appear only when needed.
Notion AI & HubSpot
Automatic linking and categorisation of information. Smart relationship tracking and predictive actions. Predictive text and recommended next steps - reducing cognitive load at every turn.
Designing an AI-powered analytics feature that transforms raw call data into actionable intelligence - helping sales and marketing teams make faster, better decisions.
ClientWaybeo (waybeo.com)
Duration1 year · 2016–2017
RoleUX Designer (Individual Contributor)
ScopeInformation Architecture, User Research, User Flows, Wireframes, Hi-Fis, Prototype
AI Insights is an AI-powered analytics feature that enhances sales and marketing by analysing customer call interactions. It provides actionable insights to improve engagement, optimise marketing efforts, and refine sales strategies - helping businesses make data-driven decisions efficiently.
The existing product lacked AI capabilities and offered only a basic analytics dashboard without deeper intelligence. The opportunity: introduce a native AI Insights tab with a customisable dashboard and recommendations that surface critical insights seamlessly.
Faster decision-making through improved data visualisation
Native AI insights - not retrofitted, built-in from day one
Actionable recommendations for sales & marketing teams
Before
The existing product - no AI, no actionable insights.
01
Understanding the user
User research · Competitive analysis · Personas · Problem statements
User Research
Stakeholder discussions and competitive benchmarking.
Stakeholder and customer discussions shaped the direction - covering the right formats and visualisations for insights, how to simplify complex data, alignment with customer workflows, and the role AI should play in improving sales efficiency.
3Competitors benchmarked - Exotel, Knowlarity, and existing Waybeo product
1Clear differentiator: native AI insights with user-friendly, actionable recommendations
5 wkEnd-to-end design cycle from research through to high-fidelity prototype
Competitive Analysis
Where competitors fall short - and where Waybeo can lead.
Pain Points
Four gaps in the existing analytics landscape.
01
Scattered & Unstructured Data
Existing platforms present raw data without clear insights, making it difficult for businesses to extract meaningful information from call interactions.
02
Lack of Actionable Insights
Competitors offer basic analytics but fail to provide AI-driven recommendations that help businesses optimise their sales and marketing strategies.
03
Complex & Non-Intuitive UI
Many analytics tools require technical expertise, making it challenging for non-technical sales and marketing users to navigate and leverage insights.
04
Inefficient Decision-Making
Without a centralised AI-powered dashboard, businesses struggle to make quick, data-driven decisions that improve engagement and conversions.
Persona
Who we designed for.
Sales Lead - Auto Dealership
Rajesh Vashi
Age 34 · Mumbai
Rajesh is a sales executive lead at an auto dealership who handles a high volume of customer calls daily. He relies on call data to track leads and close deals, but the current system gives him raw numbers with no context or follow-up guidance.
Goals
Increase sales conversions with better customer engagement
Use AI insights to understand customer behaviour and improve follow-ups
Prioritise high-potential leads for efficient selling
Frustrations
"Struggles to analyse call data manually, leading to missed opportunities"
"Wastes time sifting through unstructured data instead of selling"
"Lacks actionable insights to refine sales strategies effectively"
2×
manual analysis time wasted per week - missing high-potential leads and slowing decision-making at a critical point in the sales cycle
02
Information architecture
Four sections · Seamless navigation · AI-driven structure
Information Architecture
Structuring AI insights into four clear, navigable sections.
AI Insights structures data into four key sections: Dashboard (sales metrics overview), Potential Opportunities (lead likelihood scoring), Models (product-level insights), and Call Log (interaction-level analysis). This IA-driven approach enables seamless navigation, making AI-powered insights more accessible, actionable, and intuitive.
01
Dashboard
High-level sales metrics - total calls, AI-processed calls, sales intent score, potential opportunities breakdown, top dealers, top keywords.
02
Potential Opportunities
Lead likelihood cards (Likely / Neutral / Not Likely) with top reasons, top questions asked, and trend data for prioritising follow-up actions.
03
Models
Product-specific insights including transmission, fuel type, and colour preferences with variant-level details to inform inventory and marketing decisions.
04
Call Log
Interaction-level analysis with AI call summaries, keyword tags, and a detail overlay for reviewing individual call recordings and transcripts.
03
Starting the design
Wireframes · Draft UI · Stakeholder review
Wireframes
Low-fidelity layouts for all four sections.
Dashboard & Models
Dashboard detail
Potential Opportunities
Call Log
Draft UI
Initial designs presented to stakeholders for feedback.
Usability study · Before & after · Hi-fi prototype
Usability Study
Four key findings that shaped the final design.
01
Date Range Selection
Users found the date picker controls unclear. Simplified the filter row to make time-period selection faster and more intuitive.
02
Simplified Model Overview
The Models section contained too much data density. Reorganised the layout with clearer hierarchy and tab-based navigation by model variant.
03
Cleaner Navigation Structure
The left-hand navigation lacked clear active states and section separation. Refined with stronger visual hierarchy and consistent active indicators.
04
Call Log & Slider Presentation
The call detail overlay was difficult to scan. Enhanced the slider layout with better structured call summaries, keyword tags, and recording controls.
Before & After
How each screen evolved after usability testing.
Before
After
Before
After
Before
After
Before
After
High-Fidelity Prototype
The refined product - AI Insights live in the Waybeo platform.
Outcome
Impact and what I learned.
Happy end users
Intuitive data accessibility, improved sales conversions, and a competitive differentiation through native AI.
Scalable AI framework
The design established a foundation for future AI feature growth across the Waybeo platform.
Room to grow
Data depth, personalised insights, and better end-user guidance were identified as the next priorities for the product.
AI Insights has completely transformed how we analyse customer interactions - it’s intuitive, actionable, and has directly boosted our sales conversions.
Designing enterprise-grade scheduling, optimisation, and insight tools that balance operational complexity with everyday usability.
ClientVerint
Year2023 – 2025
RoleUX/UI Designer
ScopeUX/UI Design, Interaction Design, Accessibility Standards, User Research, Design Reviews
Overview
Enterprise complexity, human-centred design.
Verint's Workforce Management (WFM) Suite is a comprehensive enterprise solution designed to optimise scheduling, improve customer and employee experiences, enhance flexibility through mobile tools, and provide actionable insights via scorecards.
The product serves contact centres and large operations teams where scheduling errors carry significant business consequences. Every design decision had to balance the rigour demanded by enterprise users with the clarity needed for day-to-day efficiency.
Challenge
Designing for power users in high-stakes environments.
WFM users are experts in their domain - they think in shift patterns, coverage metrics, and service level agreements. The design had to honour that expertise while eliminating unnecessary friction and surfacing the right controls at the right moments.
Accessibility was a non-negotiable requirement, with compliance standards built into the design process from the outset rather than retrofitted at the end.
Process
Embedded in the product team.
Working as an embedded designer within Verint's product team, I participated in user research sessions, design reviews, and cross-functional collaboration with engineering and product management.
The iterative design process involved regular usability reviews, accessibility audits, and close collaboration with developers to ensure design intent was preserved through implementation.
Outcome
A WFM suite users trust.
The redesigned suite improved scheduling efficiency, reduced training time for new users, and achieved compliance with accessibility standards. The design system established during this project became the foundation for future WFM product development.
Designing mission-critical interfaces for A350 and A380 aircraft maintenance, and a secure authorisation platform for enterprise services.
ClientAirbus
Year2019 – 2023
RoleUX/UI Designer
ScopeUX/UI, Interaction Design, Information Architecture, Prototyping, User Testing
ArGO - Aircraft Ground Operations
Interfaces for aircraft that don’t forgive mistakes.
ArGO is a specialised application for managing the maintenance and operation of commercial aircraft on the ground and in flight - primarily for Airbus A350 and A380 models. The service is provided by Airbus to airline companies worldwide.
The stakes here are extraordinarily high. Every interface decision - a label, an information hierarchy, a confirmation flow - has safety implications. Designing in this domain required extreme rigour, deep collaboration with domain experts, and continuous testing with real maintenance engineers.
Large volumes of highly sensitive data are transferred between aircraft and ground systems through ArGO, accessible only within secured Airbus networks. Security and clarity had to coexist throughout.
Process
Deep domain immersion.
The project began with extensive field research - understanding how maintenance engineers actually work, the cognitive load they operate under, and the failure modes of existing tools. User journeys were mapped in granular detail before any interface work began.
Prototypes were tested iteratively with actual ArGO users, incorporating feedback from both technical stakeholders and end users in airline maintenance operations.
Core Elec - Authorisation Platform
Managing access at enterprise scale.
Core Elec is the authorisation platform underpinning the broader Airbus digital ecosystem - managing access to APIs, UI surfaces, and user roles across multiple interconnected products.
Designed for a large-scale solution where multiple products work in concert, the platform required a clear, consistent model for permissions, roles, and access states that technical administrators could operate confidently.
Outcome
Trusted by airlines worldwide.
The ArGO interface shipped to airline partners and has been in active use for A350 and A380 fleet management. Core Elec established a reliable access-control foundation for Airbus's growing digital product suite.
Building a mobile-first sales call management app for small and medium businesses using Material Design principles.
ClientTring Partner (tringpartner.com)
Year2015 – 2016
RoleUX Designer
ScopeUX, Material Design, Development Collaboration
Overview
Sales call management in your pocket.
Tring Partner is a mobile-based application for small and medium businesses to manage sales calls within a team or across multiple teams. The product targets the growing SMB segment that needs CRM-like capabilities without the enterprise overhead.
The design challenge was building something powerful enough to be genuinely useful for sales teams while remaining simple enough for adoption across non-technical users in small businesses.
Process
Material Design as a foundation.
The project used Material Design as the foundational design language, collaborating closely with the development team to ensure the implementation matched the intended experience.
I worked across UX flows and UI execution - designing the information architecture, interaction patterns, and visual components that made the app feel native and intuitive on Android devices.
Outcome
Streamlined team sales operations.
Tring Partner shipped with a clean, consistent Material Design implementation that gave SMB sales teams a reliable tool for call logging, team coordination, and performance tracking - all from their mobile devices.
Designing a unified super app for small and medium businesses on Android.
ClientOnne App (onne.world)
Year2016 - 2017
RoleUI/UX Designer
ScopeUI, UX, Information Architecture, Android
Overview
One super app.
Onne App is a super app for small and medium businesses, built natively for Android. The product - Onne App and Onne Business - aimed to consolidate team communication into a single, reliable mobile experience.
For SMBs operating without dedicated IT infrastructure, a communication tool needs to be intuitive from day one, reliable under real business conditions, and simple enough that the entire team adopts it.
Design Approach
Structure through information architecture.
The project required careful information architecture work to organise the app's features - messaging, calls, team management, and business tools - into a hierarchy that felt natural rather than cluttered.
The design balanced the breadth of functionality with a clean, uncluttered interface, ensuring that everyday communication tasks required minimal steps and cognitive effort.
Outcome
Communication that just works.
Onne App launched on Android with a structured, navigable interface that brought together the key communication needs of SMB teams. The information architecture established a clear foundation for future feature expansion across Onne Business.
Next Project
SalesX - Voice-First CRM.
Process & Philosophy.
Great design solutions don't exist in isolation - they are part of larger systems. Understanding the full context, the people, their needs, and the environment they operate in is where meaningful design begins.
Evidence-based design
Design decisions grounded in data, research, and rational insights rather than intuition alone. Every choice answers a question, informs a decision, or validates an assumption.
Calm technology
The best design fades into the background, letting people focus on what matters. Technology should be a quiet enabler - seamless, intuitive, and unobtrusive.
Systems thinking
Using user journeys and service blueprints to map the interconnectedness of different parts of the system, creating cohesive experiences across physical and digital touchpoints.
Experimentation
Design as exploration. Prototyping quickly, testing assumptions, and iterating based on what's learned - embracing uncertainty as a path to better solutions.
Collaboration
The best work happens at the intersection of disciplines. Facilitating workshops, bridging engineering and design, and building shared understanding across teams.
Leadership
Guiding teams through ambiguity, establishing design processes, and creating the conditions where great work can emerge. Design leadership is about enabling others.
Learning through building, failing, and simplifying what's complex.
I've spent that time learning how to navigate complexity, adapt quickly, and build with intent - largely focused on making systems easier to understand and use, while balancing business needs and user expectations.
Beyond product design, I've explored different paths early in my career - experimenting with building communication solutions, working on branding initiatives, and even attempting to create a community platform for filmmakers and reviewers. Not everything worked, but each experience shaped how I approach problems today - with curiosity, resilience, and a bias toward action.
I enjoy collaborating with people who challenge ideas and push for better outcomes. Whether it's mentoring designers, working closely with cross-functional teams, or refining workflows, I care about creating environments where good design can thrive.
Outside of work, travel plays a big role in how I see the world. I genuinely love exploring new places, experiencing different cultures, meeting people, and gaining fresh perspectives wherever I go. I've travelled to over a dozen countries so far, and I'm always looking for the next experience.