(Ala, shall we flip the HTTP protocol on its head?)
Feriwala: A feriwala is a street vendor who travels around selling wares—bringing the market to you, instead of the other way around.
Today, the internet resembles a vast, sprawling bazaar where users "wander" to find what they need—whether it's food delivery (Swiggy, Zomato), cabs (Uber, Ola), or e-commerce (Flipkart, Amazon). But what if we flipped this model? Imagine merchants proactively coming to your "doorstep" with offers, like the traditional feriwalas.
Now, imagine everyone having a personal assistant with unlimited capabilities to:
- Scan webpages.
- Record prices.
- Update shopping carts.
- Even negotiate on your behalf.
With advancements in AI, this is no longer a far-off dream. We’ve already witnessed the evolution from offline commerce to e-commerce, then to 10-minute quick commerce. The next leap could be true embedded commerce, where merchants actively ping customers with offers—selling secondary or incidental goods to someone engaged in an unrelated principal journey.
A Personal Shopping Assistant at Your Service
With an AI-powered assistant handling interactions on your behalf, this future becomes akin to having a personal concierge that:
- Knows your context.
- Understands your preferences.
- Negotiates on your behalf.
- Filters out the noise of irrelevant options.
Here’s how it could work in two simple steps:
- Users share their intents with their trusted assistant or platform.
- Sellers respond to these broadcasts in real-time.
Example:
- A user tells their assistant, "I need a laptop under ₹50,000."
- Merchants receive this "call" and ping back with tailored offers in real-time.
Flipping the HTTP Protocol
Currently, the internet operates on the HTTP protocol, invented by Tim Berners-Lee in 1989. Its three-step handshake, built on the TCP protocol (the transport layer), forms the bedrock of how information flows.
The current model is request-response based:
- The user (client) initiates a request, and the server (service provider) responds.
This works well in today’s world of ~5 billion users connecting to ~2 billion web apps (only ~400 million of which are active). The dominant platforms (Amazon, TikTok, WhatsApp, etc.) handle most of the traffic, as users manually scan a few websites to meet their needs.
But this model has its limitations. It was designed for human users navigating a handful of websites—not for a world where every user has an AI assistant capable of scanning millions of options simultaneously.
Now, imagine a new model where the HTTP protocol is flipped:
- Instead of users requesting services, sellers (merchants, or feriwalas) proactively ping users with offers tailored to their explicit or implicit needs.
This paradigm shift—from "call-OUT" to "call-IN"—has tremendous implications for e-commerce, banking, financial services, and the ad economy.
Implications for Commerce and BFSI
In its current form, embedded finance (e.g., offering insurance while booking a flight) relies on explicitly or implicitly embedding services into primary journeys. But with this flipped model:
- Secondary journeys (like getting insurance while booking a car) become proactively triggered by merchants, based on user-defined intents.
- Intent classification could range from needs → wants → desires → wishes.
Ad Economy Transformation
The traditional advertising model—built on display ads, push notifications, and enticing messaging—will have to evolve into consent-based, intent-driven messaging.
Likewise, the data economy will shift, as the first layer of consent is inherently handled (since the user initiated the intent). Merchants won’t rely on guessing user behavior but instead act on clearly defined intents.
The Role of a New Protocol
The key innovation here is a new protocol tailored for the AI-assistant paradigm. Instead of users always initiating requests, merchants could proactively engage with clients based on:
- User-defined intents.
- Context-based triggers.
In many ways, this resembles bidirectional WebSockets or Server-Sent Events (SSE) but applied to commerce at scale.
Challenges:
- Preventing spam or irrelevant pings from merchants.
- Managing the number of open connections an AI assistant can handle.
- Using ML techniques to curate merchants and offers based on user preferences.
The Vision: A New Era of Embedded Commerce
Imagine a world where:
- Commerce is personalized and proactive, not reactive.
- Financial services and e-commerce converge seamlessly into user journeys.
- The HTTP protocol evolves to handle push-based interactions.
This isn’t just a technical innovation—it’s a fundamental shift in how we interact with the digital economy.
Authored by:
Akhil Handa
Linkedin: linkedin.com\in\akhilh