Why AI in India won't look like AI in San Francisco

The tools are the same. GPT-4, Claude, Gemini — they're accessible everywhere. But the market, the constraints, and the opportunities are completely different in India.

Here's what I mean.

The infrastructure gap is the opportunity

In SF, AI is layered on top of well-functioning systems: reliable email delivery, predictable payment rails, strong B2B software culture. AI improves existing workflows.

In India, those systems are often absent or fragmented. WhatsApp is the CRM, the email, and the customer support system for millions of businesses. There's no Salesforce — there's a group chat.

This is the opportunity. AI here doesn't improve workflows. It creates them.

WhatsApp-native beats web-native

A restaurant in Bandra doesn't need a customer portal with login. They need orders taken on WhatsApp, paid via UPI, confirmed via an automated message. The "AI product" is a flow, not a dashboard.

This matters for how you build. A tool built for an Indian restaurant owner looks nothing like a tool built for a US restaurant chain. Same problem, different constraints, completely different solution shape.

Vernacular will win before English

The next 300 million Indian internet users speak Hindi, Marathi, Tamil, Telugu — not English. The current generation of AI tools is English-first by default.

The businesses that build WhatsApp agents in local languages, with local payment integrations, targeting Tier-2 and Tier-3 cities — those will win the next decade.

I'm not there yet. But it's where I'm building toward.

The practical implication

If you're building AI for Indian businesses, stop copying what you see on Product Hunt. The aesthetics, the pricing, the feature set — none of it translates directly.

Think: what does this business actually do on a Tuesday? What do they spend 3 hours on manually? That's your AI opportunity.

That's the whole point.