Language support in urgent situations
Tarjimly connects refugees, aid workers and people in crisis with translators across more than 250 languages. The app worked, but the route to help could be difficult to follow under pressure.
Users needed to start a request quickly, understand their options and stay confident while waiting. My work covered the journey from the first action through to the saved session.
Where people struggled to start a request
The research showed hesitation before starting requests, drop-off before connection and difficulty choosing between support options under stress.
I kept the useful options and changed how people reached them. The main actions needed to be clear, with other details appearing when they became relevant.
Making calls and translations easier to find
I made live interpreter support and text or document translation easier to identify at the starting point. People could choose the type of help they needed before working through the rest of the request.
The call and translation journeys used the same visual language, with their differences explained through the task.
Collecting the details needed for matching
Onboarding introduced the service. Language selection and request topics then added information needed for matching and service quality.
Where possible, the product suggested or detected language and left non-critical details until later. This reduced the number of decisions someone had to make before getting started.
Keeping requests and session history connected
I designed matching, wait status and request confirmation so users could tell that the system was working. When immediate connection was unavailable, the flow could suggest scheduling.
The journey also needed to continue through calls, chat, translation, transcripts and activity history. People should be able to return to a request without rebuilding the context.
Integrating Spark and product feedback
Spark added immediate AI help, clarification and human review. I used shared patterns so these states fitted into the wider app rather than feeling like a separate product.
Lottie covered loading, matching, call and recording states, confirmations and completion feedback. I used Jitter for motion videos and reusable assets across product, web and marketing.
Connection time and platform scale
The user and translation totals describe product scale, rather than an uplift caused by the redesign.
The redesigned experience helped people connect in under 30 seconds across language pairings and reduced drop-off in urgent requests.
What I learned about designing for urgency
This project made me more deliberate about what to ask of users under pressure. Useful information still mattered, but the product needed to collect it in an order that helped people move forward.
