There’s that moment in a cold call when the receiver is already picked up, the person is talking, and you’re nodding — but inside it’s total silence. You don’t understand a word. Especially if it’s a Brit with a strong accent on the other end, or an Indian caller who speaks very fast and very confidently.
That’s exactly what happened to me in the first weeks of working as a broker in Dubai. I’m a designer, I know how to design complex systems, I understand product — but a phone call in English with an unfamiliar accent is a whole different skill. And it doesn’t appear instantly.
You could just practice and wait. I decided to build a tool.
What is CallCoach
CallCoach is a native iOS app built with Swift, Apple Intelligence, and Live Translation.
During a call it does this in real time:
— recognizes the other person’s speech and shows the text right on the screen
— analyzes what they said
— suggests prompts — what to answer, how to continue the conversation, which question to ask
This isn’t transcription for transcription’s sake. It’s a script assistant that pops up exactly when you need it — in a live conversation, in real time, with no lag.
Why Swift and Apple Intelligence
I could have used any web stack and made this as a web app. But the web doesn’t give the speed and native integration with the microphone and system APIs that I needed.
Apple Intelligence is a powerful on-device language engine running right on the device. No cloud requests, no delay, no leakage of call data. Everything happens on the phone.
Live Translation gave me speech recognition with real accuracy even on accented English — exactly what’s needed for cold calls in multicultural Dubai.
How it changed my work
The first tests showed: when you can see the text of what the other person is saying, the anxiety fades. You stop spending energy on “what did they say” and start thinking about “what should I answer.”
That’s a fundamental shift. From survival mode — to dialogue mode.
I keep developing it. The next step is a script database for different scenarios: first call, objection, budget clarification, agreement to meet. Prompts that appear not just as transcription, but as an intelligent contextual response to a specific phrase from the other person.
Why I’m sharing this
Because CallCoach is an example of how I think about products.
Not “which technology should I try,” but “which problem should I solve.” The problem was specific and personal. The solution is native, fast, and without unnecessary dependencies. A tool is born from real pain, not from the desire to build something.
A designer who can write code is not a developer. But sometimes that’s exactly what’s needed so the idea doesn’t die in the implementation queue.