Role
Product and UX design
Scope
Type
Self initiated
Year
2026
The problem
The app shows status when the passenger needs a decision. So they leave and check somewhere else.
The approach
One hub that changes state across the journey, always answering a single question: what do I do now.
The outcome
Six states covering the calm path and both failures, with honest odds, a countdown, and a real backup plan.
The problem
It starts at 5:40 PM
Leela and Vishal are booked on the 12472 Swaraj Express, New Delhi to Nagda, leaving at 21:40. They paid ₹1,401. They have a PNR. What they do not have is a seat.
They are waitlisted. Somewhere around 5:40 PM today a chart gets prepared and a computer decides whether they travel. Between paying and that moment, the app has almost nothing to say. It shows a line that reads Chart not prepared and a status that has not moved in hours.
So they do what everyone does. They refresh. They open a third party app to check the odds. They google whether waitlist 23 usually clears.
They are not looking for information. They are looking for a decision they can make.
Status, not guidance. The confidence percentage shown at search time disappears here, and the safety net only promises money back.

Before · the current screen
What the research showed
The detail that settled the direction
There is an entire industry of apps built for one job: predicting whether your train ticket will confirm. ConfirmTkt, RailYatri, RailMitra, and more. Millions of downloads between them. They exist because people do not trust the raw status the official app gives them. Someone has already booked their ticket, and they still go somewhere else to find out what it means.
That is not a feature gap. That is a trust gap.
77%
of trains on time, and on time officially means within fifteen minutes
4 hrs
before departure, the chart decides whether you travel at all
0
guidance offered in the app during that entire window
The safety nets that do exist are quiet or hidden. Vikalp, the alternate train scheme, has to be opted into at booking. The connecting train protection returns money and nothing else. The app is not short on data. It is short on meaning.
The reframe
The system

01 · WAITING · Waitlisted and RAC, before the chart
The wait
The hardest screen, because nothing is happening yet and that is exactly the problem. Three things went in.
Honest odds per passenger. Leela is RAC 5, near certain, green. Vishal is waitlist 24 at 70 percent, blue, moving. Both show movement, because a number that visibly changes tells you the app is watching for you.
A countdown to the decision. Seats decided around 5:40 PM today. Naming it turns open ended dread into a finish line.
A backup held before the outcome. Not a refund promise. A real plan B, locked in while hope is still alive.

Nowhere does it say you will get a seat. A waitlist is never certain, and a promise you cannot keep costs more trust than silence does.
01a · The lock, and what it actually costs
Asking someone to commit to a backup while they are still hoping is a big ask, so the sheet leads with Nothing is charged now. Here is exactly what you are locking. Three plain lines follow: one Sleeper seat held on the 12138 for both passengers, triggers only if the chart does not confirm you at 5:40 PM, pay only if it triggers. Then the escape hatch, you can change or cancel the backup any time before the chart prepares. Every objection answered before it is raised.


02 · CONFIRMED · Chart cleared, berths allotted
The good outcome still needs a screen
Most redesigns treat success as the state that needs no work. It does. The moment the chart clears, the passenger’s question changes completely, from will I travel to where exactly do I stand on the platform.
So the screen changes what it emphasises. The odds and the countdown disappear entirely, because they are now answered. The berth becomes the largest thing on each passenger row, S6-34 lower and S6-37 middle, because that is the one piece of information a confirmed passenger is actually looking for. Under it, boarding at New Delhi, Platform 8, departs 9:40 PM, boarding opens 9:10 PM.
The primary action shifts from locking a backup to tracking the live journey, with setting an arrival alarm underneath. The trip is calm now, so the screen gets calm with it.
This is what one hub, many states really means. Nothing moved to a different page. The same surface stopped answering a question that no longer exists and started answering the next one.


03 · DELAY AND PLATFORM · Running late, platform moved
The day itself
The train is confirmed, running 40 minutes late, and the platform has moved from 8 to 12. Most apps flash a red banner and stop. This screen gives the three things you actually need standing in a station: the new platform, where your coach stops on it, and the revised arrival. Then the choices, with one clearly recommended. Not a menu.
03a · BOARDING CHECK · The train has just departed
The moment the train leaves
Missing your train is the worst failure in Indian rail travel, and no app catches it. So the moment 12472 actually departs, not the scheduled time, the live departure, a gentle check appears. Are you on board?
Yes opens the live journey. No does not hand you a helpline number. It opens the same rescue engine the failed chart uses.
This one sheet is the hinge of the whole system. It is the only place the app asks the user something instead of telling them, and it is what makes the difference between a tracker and something that is actually looking out for you.
Two things I got wrong first
I originally triggered it five minutes after the scheduled departure, which is wrong when the average delay runs past half an hour and the train is often still sitting there. And I made it a hard gate, which nags the 95 percent of people who are perfectly fine. It became a soft card on real departure, defaulting to the good path, with a quiet line underneath: still boarding? You can answer in a minute.

04 · ON BOARD TRACKER · Travelling, train on time
Eight hours of nothing to do
The passenger said yes. Now they are on a train for eight hours overnight, and the only question left is when do I need to be awake.
The screen becomes a station by station progress line. Departed New Delhi on time at 9:40 PM, currently near Mathura Junction, next stop Agra around 11:10 PM, then Agra Cantt, then arrives Nagda around 5:50 AM tomorrow. Under it, one commitment: we will wake you 30 minutes before your stop.
This is the state where a flight style boarding tracker would have been the obvious pattern to reach for, and it would have been wrong. Trains do not board on a gate clock and they do not have gate numbers. Real station progress is what a rail passenger actually tracks, so that is what the screen shows.

05 · MISSED TRAIN RESCUE · The train left without you
They said no
The screen says it plainly. You missed the Swaraj Express. No softening, no apology copy. Then it immediately does something useful, because the situation is not actually hopeless and most people do not know that.
The train is near Ghaziabad. It reaches Agra at 11:10 PM, tagged Catchable. Your ticket is valid till Agra. You have about 1h 20m to get there, and if you board there your original ticket still works. That is the recommended option, and it is recommended because it costs nothing.
Below it, ranked by cost, the 12920 Malwa Express at 11:30 PM with Sleeper seats at ₹430, or a bus at 12:15 AM at ₹620. The primary action is get directions to Agra, with filing for a refund kept available underneath rather than presented as the answer.
Missing a train is stressful. We will help you get moving.


06 · CHART FAILED RESCUE · The waitlist did not clear
The chart did not clear
The failure state most redesigns skip. The screen says it plainly, no softening. Then, immediately, the refund is already moving. And under that, three ranked ways to still reach Nagda tonight, starting with the seat that was held four hours ago, one tap to take it.
That is why the backup exists. If you pre committed a plan B while things were still uncertain, the failure stops being a scramble and becomes a handoff.
One tap and you are booked. Your refund comes separately. The two things are deliberately decoupled, because waiting for money to arrive before you can travel again is the exact anxiety the whole redesign exists to remove.

Add on
Making the hub find the user
Everything above assumes the passenger opens My Trips and taps into their booking. An anxious one will. A calm one will not, and the calm one is exactly who needs to know the platform moved.
So the hub also surfaces itself. A live trip strip sits on the ixigo home screen showing the single most urgent thing about the current state, here Delayed by 40 minutes, next Rewari, under the train name. Tapping it opens whichever state applies, straight into the screen from the sections above.
It deliberately reuses the Flight Tracker pattern the app already has on that home screen, so it is a placement decision rather than a new component. The urgent state finds the user the moment the app opens instead of waiting two taps deep inside My Trips.
Labelled clearly as an add on. It is not part of the core post booking flow and the redesign stands without it.
Prototype not loading? Open it in a new tab→
What it cost
Trade offs
01
Hierarchy over completeness
The main trade off was hierarchy. The decision leads, so the ticket, the payment and the utilities move below it. That is the right call for an anxious waitlisted passenger, who cares about their seat first and the paperwork second. It is arguably the wrong call once the trip is calm and confirmed. I accepted that because the anxious state is where the app is failing hardest right now.
02
Honesty over comfort
The second was honesty over comfort. Showing 70 percent means showing that 30 percent exists. That is a harder screen to look at, and it is the only version that survives being wrong.
Assumptions and scope
What I took as given
01
This is a single direct train, so missed connections do not apply. I swapped that case for the missed train check, which is the real equivalent failure on a direct journey.
02
The held backup builds on ixigo’s existing Alternate Travel Plan, reframed from a passive refund into an active seat hold. The honest version is Vikalp made usable, not magic inventory.
03
Confidence percentages and live positions are illustrative, and assume access to IRCTC status and running data.
How I used AI
A thinking partner, not the designer
It moved fast through the searchable and mechanical parts: pulling and sanity checking rail data, mapping the full space of post booking states so I could see the system before committing to screens, generating options to react against, and drafting copy I then cut back.
What I rejected matters more than what I accepted.
Rejected
Over reassuring copy, because a broken promise is worse than an honest number.
Rejected
A missed connection scenario it proposed. This is a direct train, so inventing a connection would have made the screen lie about the trip.
Rejected
More dashboards, because the point was never to show more data.
Rejected
A flight style boarding tracker, because trains do not board on a gate clock and those states meant nothing to a real passenger.
That last one is the useful lesson. The AI gave me a pattern that looked right and was wrong for the domain. Knowing which is which is still the job.
What the design review changed
A screen can be strategically right and operationally useless
Practising designers reviewed the first build and pushed back on the information architecture. Their point was that a post booking screen has four things a passenger reaches for, the PNR, the e ticket and QR, the train details, and the ticket details, and the first version had buried two of them.
01
PNR was styled as a subtitle in the header. It is an identifier a passenger reads aloud to a TTE and types into status checks, so it moved onto the journey card as a labelled field with a copy action.
02
E ticket and Scan QR sat two scrolls down. Ticket access is time critical at a gate, so both moved into the first fold, directly under the journey card.
03
Seat numbers on the confirmed screen carried no emphasis. The berth is now the visual anchor of each passenger row.
04
Platform numbers were showing on the waiting screen before a platform is ever assigned. They now appear only once the information genuinely exists.
The designers read it as users. I had been reading it as an argument.
I had ranked the screen by importance to my argument rather than by when the user needs each thing. Urgency and importance are different axes, and sorting by the wrong one is what buried the essentials.
More work
Check out more work
Twelve · Plans in years, not months
Google Pay Split and Request
Basil · Product page set builder concept



