Service design Strategy X Behaviour change • project 2024
Turning hesitant bystanders into confident first responders.
ROLE
Service Designer
& Behaviour Science
TIMELINE
Jan 2024 - August 2024
TEAM
4 Designers
SKILLS
Service Design
Ethnography
Prototyping
Behaviour Studies
Entrepreneurship
Problem
Who saves you in the first hour?
In Delhi, half the people who die in road accidents could have lived if someone nearby had helped in time. The people standing closest are rickshaw pullers, delivery riders, and street vendors who spend their working lives on the road. Most of them want to help. Almost none of them do.
Our research found the reason sitting outside the accident itself, inside the systems that respond to it.
74% won't step forward
Three in four Indian bystanders refuse to help road accident victims. Fear of police harassment, court hassles, and being blamed for the death outranks every other reason.
78% From low income backgrounds stay back
The people most likely to witness an accident are the least likely to assist, because past mistreatment by legal and medical institutions taught them the cost of helping.
Willing but unequipped
Those who overcome the fear often lack first aid knowledge, a first aid kit, or a vehicle. Good intentions stall at the moment of action.
.
Why It Mattered
India holds 1% of the world’s vehicles and the highest number of accident deaths on earth.
Delhi recorded 5,387 road accidents in 2022, with 1,412 of them fatal. More than half of those killed were pedestrians, two-wheeler riders, and other vulnerable road users.
.
Research
Tackled via 6 different methods
We combined desk research, on-site interviews, expert consultation, hospital visits, and cultural probes to understand the full system around a road accident, from the moment of impact to the police station weeks later.
Desk Research
Insight: The barrier is legal, not logistical
We analysed national accident data and SaveLIFE Foundation’s study on bystander care. The leading reasons for inaction were fear of police harassment (33.1%), avoiding court hassles (28.5%), and fear of blame if the patient dies (9.7%).
Actionable cue
We decided to treat the medicolegal process as the primary design target.
.
On-Site Interviews at Accident Hotspots
Insight: Time is a livelihood cost
We spoke with rickshaw drivers, delivery riders, and street vendors who witness accidents regularly. One delivery rider told us his customer cares only that the order arrives on time. A rickshaw driver said people don’t have time for rounds of the police station and hospital.
Actionable cue
We decided that helping must take minutes, not days, and must not put a worker’s income at risk.
.
Expert Interviews, SaveLIFE Foundation
Insight: The law exists, the trust doesn't
The NGO behind India’s 2016 Good Samaritan Law explained that legal protection alone changed little. Good Samaritans are still blamed, still questioned, still detained.
Actionable cue
We decided to design for perceived safety, which is the thing that actually governs behaviour.
.
Hospital Visits: Doctors, Nurses, EMTs, Ambulance Staff
Insight: Information arrives too late
Doctors told us they judge treatment by the mechanism of the accident, the impact, and the primary injuries. That information usually arrives as a fragmented verbal account from a frightened bystander.
Actionable cue
We decided the bystander’s observations should travel to the hospital as structured data.
.
CULTURAL PROBING - Testing Biases
Insight: Help follows familiarity
We printed cards showing people of different genders, ages, occupations, and backgrounds, then asked participants who they would help first. Auto drivers chose other auto drivers. Almost everyone chose the poorest-looking girl, assuming no one else would. Well-dressed people were chosen last.
Actionable cue
We decided the service must dispatch help on need, removing individual judgement from the equation.
.
CULTURAL PROBING - Testing Perceptions
Insight: Media writes the rules people follow
We mixed real and fake news headlines in a familiar visual format. 78% believed a false headline about police taking bribes from a Good Samaritan. 93% believed a false headline about repeated court visits. 72% believed a true headline about government rewards was fake.
Actionable cue
We decided that changing perception is as much the product as changing procedure.
.
Sensemaking
We mapped every actor whose decisions shape whether a victim survives: bystanders, police, ambulance crews, hospital front desks, trauma doctors, private hospital management, employers, delivery platforms, families, NGOs, and news media.
We then ran two journey maps side by side, the ideal first responder journey against the real one lived by someone who had helped. The gap told the whole story.
Ideal: Call for help, give first aid, hand over to a trained ambulance team, go home. Police make contact only if the bystander volunteers as a witness.
Real: Carry the victim in your own vehicle. Tie a cloth around the wound. Explain to the police on scene that you did not cause the crash. Hand over your details. Watch a medicolegal case get filed in your name. Return to the station repeatedly to give statements. Learn to wait for someone else next time.
.
Through the COM-B behaviour model, we traced the undesired behaviour to three gaps:
.
The Leverage Point
We plotted a vicious circle that keeps bystanders trapped.
Bad experiences create bad perceptions. Bad perceptions stop people from helping. No one helping means no good experiences get created, and the cycle holds.
One intervention breaks it: the moment when nobody comes forward.
A single positive experience of helping spreads through word of mouth and reshapes what a community believes is possible. We designed for that first domino.
.
Ideating & Testing
We ran Top 5, SCAMPER, and Service Flip to generate ideas across five need clusters.
Which are overcoming hesitation, protection and clarification, community response, education and awareness, and on-site support. Desk research filtered them for feasibility.
Validating with Users: We surveyed and interviewed potential bystanders.
89.7% said a first aid trained volunteer arriving before the ambulance would help.
67.9% stayed neutral on monetary rewards, confirming what our interviews suggested.
51.9% wanted free first aid training.
50.9% said anonymous reporting would motivate them to act.
.
Prioritising Features
A Kano Model analysis separated must-haves from attractive extras and shaped three concepts. We scored each against our five target needs.
Matrix
Concept 2, built on anonymous reporting and dispatched community volunteers, met all five.
.
Concept Feedback
We took Concept 2 back to low-income bystanders.
What we learned
They rejected cash rewards, calling the act humanitarian, while still feeling unrewarded for the time and risk involved. They disliked performing first aid over a phone call and welcomed the ability to report anonymously.
Actionable cue
We removed cash incentives and replaced them with healthcare benefits. We moved first aid from the untrained caller to a trained responder on the ground.
.
Solution
Parakram (पराक्रम); meaning valor, bravery, strength, might, or courage.
Parakram lets bystanders call for help anonymously and brings trained citizen responders to the scene before the ambulance arrives. Victims get care inside the golden hour. Bystanders get protection from the process that once punished them.
WatCH THE SERVICE STORY HERE LINK
.
Next steps
What about accidents with no bystanders around?
In areas with fewer people around to respond to accidents, the Parakram device uses crash sensors, GPS, IoT connectivity, seat-occupancy detection, and a siren to detect crashes in real time.
How it works
Its SOS system immediately alerts the nearest hospital, police station, and registered emergency contacts, while also notifying nearby Parakramveers, first-aid-trained volunteers based at local medical stores who can quickly reach the scene and provide immediate assistance.
.
IMpact
Validation in the market
A year after we conceptualised Parakram, Zomato and Blinkit launched a near-identical service: 10-minute ambulance response built on their existing delivery fleet and operational footprint.
????
The resemblance is coincidental, and that is exactly what makes it useful. Two independent teams, looking at the same gap in India's emergency response system, arrived at the same answer: the fastest path to a victim runs through people already on the road.
What we could only propose as a student concept, a company with fleet infrastructure and capital took to market. The model works. It needed resources, not a different idea.
We proved the logic. They proved the feasibility.
saraSanwal2026™