B2B | SaaS | 2024
B2B | SaaS | 2024
Building a Clinic Management App
for a Seed-Stage Startup
Health Management Information tool (HMIS) is a budget-friendly built for India's standalone clinics, pharmacies, and labs — the segment underserved by enterprise players like Practo and MocDoc. I joined as the sole product designer, working directly with the CEO, and turned the prototype into the company's primary acquisition tool.
Built from scratch
3 Personas
Ship fast, design right
2024 – 2025
MY ROLE
Lead Product Designer (solo)
TEAM
1 Product Designer (Me)
2 Engineers
Founder CEO
TIMELINE
4 months,
phased shipping
STAKEHOLDERS
VP Design
Product Head
STATUS
Live in Production
01 — CONTEXT
A pre-seed startup pitching
to a market the big players ignored.
India has 1.5M+ standalone clinics still running on paper registers, sticky notes, and WhatsApp queues. The market leaders — Practo, HealthPlix, MocDoc — chase hospitals and chains with budgets to match. Udeti positioned itself as the affordable, ABDM-compliant HMIS for the clinic owner who has none of that infrastructure. The CEO was actively pitching to clinics in Bengaluru and Kerala. My job was to design the product that gave him something real to show.
"Design decisions backed by reasoning will convince me. No need for detailed user research — use your instinct and experience to master the product."
— Brief from Anurag Aggarwal, CEO, at project kickoff
01 —WHY THE CEO BET ON INSTIN
No research budget. No design team. One designer with a track record.
Pre-seed reality: no money for 8-week discovery sprints, no time for formal usability research. The CEO needed someone who could make defensible decisions on day one — and ship.
WHAT I BROUGHT INTO THE ROOM
CleverTap SaaS depth. Years on a complex B2B product with deeply layered permissions, multi-role workflows, and conditional UI states — the exact same problem shape as a clinic OS.
Mysphere referral. Anurag was introduced to me by a freelance client (Mysphere) who'd worked with me directly — a trusted source vouching for speed and output quality.
0→1 mobile experience. Comfort with shipping in phases, defending decisions verbally, and working at startup velocity without losing craft.
HOW THAT TRANSLATED INTO THE PROCESS
Mapped 3 personas in FigJam with the CEO — Receptionist, Doctor, Admin — and built the IA around their actual sequence of work, not the app's modules.
Skipped formal research. Used SaaS pattern intuition + clinic-floor observation as the source of truth. Every decision came with a reason I could defend out loud.
Built a Figma component library first, then assembled screens — so the CEO could see a coherent system within weeks, not months.
03 — HOW THE WORK SHIPPED
4 months. Phased delivery.
Receptionist first — because she's the bottleneck.
The product wasn't designed front-to-back. It was sequenced by who has the most pain. The receptionist is the operational nucleus of any clinic — get her flow right and the rest follows. Each phase delivered key screens to the CEO for approval, then detailed workflows + dev-ready specs to engineering, all within weeks.
Phase 01 Week 01
Receptionist
Appointments, patient lookup, booking flow, walk-ins, token system
PHASE 2 · WEEKS 7–10
Admin
Clinic profile, staff management, role access, finance, templates
PHASE 3 · WEEKS 11–14
Doctor
Patient records, diagnosis, Rx templates, certificate, referrals
PHASE 4 · WEEKS 15–16
ABHA + Billing
ABDM compliance flow, auto-receipt generation, share-to-patient
04 - SIGNATURE DESIGN DECISIONS
Three calls I'd defend in any interview.
Pre-seed pace meant most decisions had to be made fast and stick. These three shaped the product more than anything else — and each came from a specific reasoning I could articulate to the CEO, to engineering, and to a clinic owner mid-demo.
Decision 01
Bottom-sheet booking, modelled on Google Calendar — but tighter.
Receptionists book appointments dozens of times a day under queue pressure. The pattern had to feel like Google Calendar's quick-add — open a sheet, pick the slot, confirm, done — not a full-screen modal that breaks flow.
Why bottom sheet, not a separate page: the receptionist needs to stay anchored to the day timeline while booking. The sheet slides up, captures the booking, dismisses — the timeline view never disappears. One mental model, no context loss.
The 10-second benchmark wasn't aspirational — it was the design constraint. If booking took longer, staff would fall back to the paper register. I measured every iteration against it.
Decision 02
The token number system — discipline by design.
Indian clinic queues are chaotic. Patients constantly ask "Can I go in next? Mine's urgent." Receptionists become traffic cops. I designed around an existing cultural pattern — the token number — and made it the spine of the entire appointment system.
Every appointment card shows a token. Patients see their own token + the current one being seen. The clinic gains predictability and order, the receptionist stops being interrupted, and the doctor's queue moves cleanly.
This wasn't a UI flourish. It was an interaction design decision that respected how Indian clinics already work — and removed the social friction without forcing behaviour change.
Decision 03
Role-based access wired into the appointment status chain.
Three roles. Three home screens. But the harder design problem was chaining access to the appointment's status. A receptionist can edit and cancel a booked appointment, but the moment she clicks "Check-In", control passes to the doctor. The receptionist literally loses write access until the consultation is complete.
This wasn't permissions for permissions' sake. It mapped the real-world handover: front desk → consultation room → billing counter. The UI physically reflects who's in control at each stage, with Admin able to configure access rules per status from settings.
And — critically — every destructive action (check-in, cancel) has an undo. Misclicks under pressure are inevitable; the system absorbs them gracefully.
"Every time I switch to the email editor it feels like I'm leaving CleverTap. The UI completely changes — fonts, colors, behavior. It breaks my flow every single time."
Growth Manager
Fintech brand (8K MAU)
"Every time I switch to the email editor it feels like I'm leaving CleverTap. The UI completely changes — fonts, colors, behavior. It breaks my flow every single time."
Growth Manager
Fintech brand (8K MAU)
I spend more time fighting the editor than writing the actual email. That can't be right."
CRM Manager
E-commerce · 2.4M MAU
"Our designer refuses to use CleverTap for email. She exports from Figma and drops in an image. We've just accepted it."
Head of Marketing
Fintech · 180K users
"We moved 40% of our email campaigns to Klaviyo because their editor actually works the way we think. We kept CleverTap for push and in-app."
CRM Manager
CRM Director, retail brand (ARPU $140K/year)
"The moment I click into the email editor it feels like I'm in a different product from 2015. Everything else in CleverTap feels modern."
Growth Lead
Travel brand · 900K MAU
"We switched our promotional emails to Klaviyo six months ago. We only kept CleverTap for push. The editor was the reason."
CRM Director
Retail · $140K ARR
"I sent a broken email to 80,000 users last quarter because I couldn't preview it properly on mobile before sending."
Email Marketer
Media company · 80K subscribers
"Every Monday I rebuild the same header and footer. I've done it maybe 200 times now. There's no way to save it."
Marketing Executive
SaaS · 45K users
"Loading the editor is the longest part of my workflow. I make a coffee and come back."
Performance Marketer
Gaming · 3.1M MAU
"Our brand guidelines say one thing. The editor only lets me do another. So our emails never look on-brand."
Marketing Designer
Consumer app · 620K users
"I stopped checking dark mode because it takes too long. I just hope it looks okay on my users' phones."
Retention Specialist
Food delivery · 1.2M MAU
"We evaluated moving to Braze just because of the email editor. The rest of CleverTap is great — this one thing was that painful."
VP Marketing
Enterprise SaaS · $280K ARR
05 CRAFT DETAIL
The "+" button that knows where it is.
A small but specific decision: the floating action button is context-aware. Same icon, same position — different action depending on screen.
One button. Four behaviours.
On Appointments → opens the booking sheet. On Patient List → adds a new patient. On Doctors → adds a new doctor profile. On Billing → creates a manual invoice. The icon never moves. The label changes. Users learn one gesture instead of four.
06- TENSIONS WITH THE CEO
Where I had to push back — and where I had to follow through.
Working solo with a founder means you're the design voice in every product debate. Two specific moments shaped the product:
07- THE PITCH
What the CEO showed clinic owners — and what closed the deal.
The Figma prototype became Udeti's primary acquisition tool. Anurag walked clinic owners and doctors through specific moments in the product that proved it understood their day.
08- OUTCOMES
What 4 months produced.
09- REFLECTION
What I'd carry into the next role.
A six-phase process with checkpoint reviews at each stage with Geetaa Bhatt (VP Design), the PM, and Anand Jain (Product Head). The process was iterative — but each phase had a clear output before the next one began.
01 — CONTEXT
A pre-seed startup pitching
to a market the big players ignored.
India has 1.5M+ standalone clinics still running on paper registers, sticky notes, and WhatsApp queues. The market leaders — Practo, HealthPlix, MocDoc — chase hospitals and chains with budgets to match. Udeti positioned itself as the affordable, ABDM-compliant HMIS for the clinic owner who has none of that infrastructure. The CEO was actively pitching to clinics in Bengaluru and Kerala. My job was to design the product that gave him something real to show.
"Design decisions backed by reasoning will convince me. No need for detailed user research — use your instinct and experience to master the product."
— Brief from Anurag Aggarwal, CEO, at project kickoff
01 —WHY THE CEO BET ON INSTIN
No research budget. No design team. One designer with a track record.
Pre-seed reality: no money for 8-week discovery sprints, no time for formal usability research. The CEO needed someone who could make defensible decisions on day one — and ship.
WHAT I BROUGHT INTO THE ROOM
CleverTap SaaS depth. Years on a complex B2B product with deeply layered permissions, multi-role workflows, and conditional UI states — the exact same problem shape as a clinic OS.
Mysphere referral. Anurag was introduced to me by a freelance client (Mysphere) who'd worked with me directly — a trusted source vouching for speed and output quality.
0→1 mobile experience. Comfort with shipping in phases, defending decisions verbally, and working at startup velocity without losing craft.
HOW THAT TRANSLATED INTO THE PROCESS
Mapped 3 personas in FigJam with the CEO — Receptionist, Doctor, Admin — and built the IA around their actual sequence of work, not the app's modules.
Skipped formal research. Used SaaS pattern intuition + clinic-floor observation as the source of truth. Every decision came with a reason I could defend out loud.
Built a Figma component library first, then assembled screens — so the CEO could see a coherent system within weeks, not months.
03 — HOW THE WORK SHIPPED
4 months. Phased delivery.
Receptionist first — because she's the bottleneck.
The product wasn't designed front-to-back. It was sequenced by who has the most pain. The receptionist is the operational nucleus of any clinic — get her flow right and the rest follows. Each phase delivered key screens to the CEO for approval, then detailed workflows + dev-ready specs to engineering, all within weeks.
Phase 01 Week 01
Receptionist
Appointments, patient lookup, booking flow, walk-ins, token system
PHASE 2 · WEEKS 7–10
Admin
Clinic profile, staff management, role access, finance, templates
PHASE 3 · WEEKS 11–14
Doctor
Patient records, diagnosis, Rx templates, certificate, referrals
PHASE 4 · WEEKS 15–16
ABHA + Billing
ABDM compliance flow, auto-receipt generation, share-to-patient
04 - SIGNATURE DESIGN DECISIONS
Three calls I'd defend in any interview.
Pre-seed pace meant most decisions had to be made fast and stick. These three shaped the product more than anything else — and each came from a specific reasoning I could articulate to the CEO, to engineering, and to a clinic owner mid-demo.
Decision 01
Bottom-sheet booking, modelled on Google Calendar — but tighter.
Receptionists book appointments dozens of times a day under queue pressure. The pattern had to feel like Google Calendar's quick-add — open a sheet, pick the slot, confirm, done — not a full-screen modal that breaks flow.
Why bottom sheet, not a separate page: the receptionist needs to stay anchored to the day timeline while booking. The sheet slides up, captures the booking, dismisses — the timeline view never disappears. One mental model, no context loss.
The 10-second benchmark wasn't aspirational — it was the design constraint. If booking took longer, staff would fall back to the paper register. I measured every iteration against it.
Decision 02
The token number system — discipline by design.
Indian clinic queues are chaotic. Patients constantly ask "Can I go in next? Mine's urgent." Receptionists become traffic cops. I designed around an existing cultural pattern — the token number — and made it the spine of the entire appointment system.
Every appointment card shows a token. Patients see their own token + the current one being seen. The clinic gains predictability and order, the receptionist stops being interrupted, and the doctor's queue moves cleanly.
This wasn't a UI flourish. It was an interaction design decision that respected how Indian clinics already work — and removed the social friction without forcing behaviour change.
Decision 03
Role-based access wired into the appointment status chain.
Three roles. Three home screens. But the harder design problem was chaining access to the appointment's status. A receptionist can edit and cancel a booked appointment, but the moment she clicks "Check-In", control passes to the doctor. The receptionist literally loses write access until the consultation is complete.
This wasn't permissions for permissions' sake. It mapped the real-world handover: front desk → consultation room → billing counter. The UI physically reflects who's in control at each stage, with Admin able to configure access rules per status from settings.
And — critically — every destructive action (check-in, cancel) has an undo. Misclicks under pressure are inevitable; the system absorbs them gracefully.
"Every time I switch to the email editor it feels like I'm leaving CleverTap. The UI completely changes — fonts, colors, behavior. It breaks my flow every single time."
Growth Manager
Fintech brand (8K MAU)
"Every time I switch to the email editor it feels like I'm leaving CleverTap. The UI completely changes — fonts, colors, behavior. It breaks my flow every single time."
Growth Manager
Fintech brand (8K MAU)
I spend more time fighting the editor than writing the actual email. That can't be right."
CRM Manager
E-commerce · 2.4M MAU
"Our designer refuses to use CleverTap for email. She exports from Figma and drops in an image. We've just accepted it."
Head of Marketing
Fintech · 180K users
"We moved 40% of our email campaigns to Klaviyo because their editor actually works the way we think. We kept CleverTap for push and in-app."
CRM Manager
CRM Director, retail brand (ARPU $140K/year)
"The moment I click into the email editor it feels like I'm in a different product from 2015. Everything else in CleverTap feels modern."
Growth Lead
Travel brand · 900K MAU
"We switched our promotional emails to Klaviyo six months ago. We only kept CleverTap for push. The editor was the reason."
CRM Director
Retail · $140K ARR
"I sent a broken email to 80,000 users last quarter because I couldn't preview it properly on mobile before sending."
Email Marketer
Media company · 80K subscribers
"Every Monday I rebuild the same header and footer. I've done it maybe 200 times now. There's no way to save it."
Marketing Executive
SaaS · 45K users
"Loading the editor is the longest part of my workflow. I make a coffee and come back."
Performance Marketer
Gaming · 3.1M MAU
"Our brand guidelines say one thing. The editor only lets me do another. So our emails never look on-brand."
Marketing Designer
Consumer app · 620K users
"I stopped checking dark mode because it takes too long. I just hope it looks okay on my users' phones."
Retention Specialist
Food delivery · 1.2M MAU
"We evaluated moving to Braze just because of the email editor. The rest of CleverTap is great — this one thing was that painful."
VP Marketing
Enterprise SaaS · $280K ARR
05 CRAFT DETAIL
The "+" button that knows where it is.
A small but specific decision: the floating action button is context-aware. Same icon, same position — different action depending on screen.
One button. Four behaviours.
On Appointments → opens the booking sheet. On Patient List → adds a new patient. On Doctors → adds a new doctor profile. On Billing → creates a manual invoice. The icon never moves. The label changes. Users learn one gesture instead of four.
06- TENSIONS WITH THE CEO
Where I had to push back — and where I had to follow through.
Working solo with a founder means you're the design voice in every product debate. Two specific moments shaped the product:
07- THE PITCH
What the CEO showed clinic owners — and what closed the deal.
The Figma prototype became Udeti's primary acquisition tool. Anurag walked clinic owners and doctors through specific moments in the product that proved it understood their day.
08- OUTCOMES
What 4 months produced.
09- REFLECTION
What I'd carry into the next role.
A six-phase process with checkpoint reviews at each stage with Geetaa Bhatt (VP Design), the PM, and Anand Jain (Product Head). The process was iterative — but each phase had a clear output before the next one began.
