How to make AI accessible to the Billions

A few months ago, I was on a call with my dad.

He lives thousands of miles away, in a part of the world where technology didn’t evolve the same way it did in the U.S. There are no dashboards in his daily life. No productivity apps. No “AI tools.”

But there is WhatsApp.

It’s where he talks to family, reads forwarded news, asks for advice, and makes everyday decisions. For him — and for most people around him — WhatsApp isn’t just a messaging app.

It is the internet.

That’s when something clicked for me.

We talk a lot about how AI is changing the world. But most of what we’re building assumes people will come to us — download a new app, learn a new interface, type in English, adapt to our systems.

My dad won’t do that.

Not because he can’t — but because he shouldn’t have to.

So I tried something different.

I built a simple AI assistant that lives entirely inside WhatsApp. No login, no onboarding, no learning curve. Just a conversation — in Bangla, the language he’s comfortable with.

Now, every morning, he can ask for:
– a quick summary of the day’s top news
– simple wellness tips
– help with everyday decisions, like what to buy or how to plan something

And he doesn’t think of it as “using AI.”

He just texts.

What started as something small for my dad made me realize something much bigger:

Over 3 billion people use WhatsApp globally. For many of them, it’s their primary — sometimes only — interface to the digital world.

But AI hasn’t met them there yet.

We building powerful tools, but often for the same kind of user — English-speaking, app-native, early adopters. But there are more than hundreds of languages that are usually left behind as early adopters.

Umar exists because I believe the next wave of AI won’t be won by better interfaces.

It will be won by meeting people where they already are.