Chatting on the internet sure has come a long way. Remember the good ole days in the ’90s and 2000s, when users would log in to AIM and overuse basic emojis, lyrics to sad songs, and lots of BRBs? With the ascendance of AI, the internet is now shifting into a new era, which some have dubbed the “conversational internet.” Champions of this newer way of being online say it’s intended to be more intelligent, interactive, and convenient. And Gupshup CEO Beerud Sheth is betting it’s soon going to “take care of everything [you] need.”
Sheth co-founded Gupshup in 2004. The AI-powered platform allows companies to create and deploy chatbots via its chatbot studio, managing over 10 billion messages each month and powering six out of 10 B2C chats globally, according to its website. IT Brew recently caught up with Sheth to learn more about the platform and find out why he thinks messaging apps are the new browsers.
On the changing landscape of the internet
“So, firstly, our vision is that unlike the US tech ecosystem, which is heavily focused on the traditional internet, as you know it—web pages, browsers, and things like that—the other half of the world is growing in the tech ecosystem; is evolving in a very different way,” Sheth told IT Brew. “It’s all mobile first. Every user is using messaging apps on those mobile devices. And there’s a whole different kind of internet emerging, let’s call it the conversational internet.”
On how this will affect IT professionals
As conversational interfaces and messaging platforms become the primary digital mediums, existing IT structures will need to adapt, according to Sheth. “As companies adopt chatbots, IT teams will be responsible for the underlying messaging infrastructure and integration complexities,” Sheth wrote in a follow-up email.
“There will be increased demand for IT skills around conversational AI, natural language processing (NLP), and generative machine learning models,” Sheth added. “Building, deploying, and maintaining intelligent chatbots requires data pipelines, DevOps orchestration, and continuous learning—all ripe areas for IT specialization.”
Top insights for IT pros
From cybersecurity and big data to cloud computing, IT Brew covers the latest trends shaping business tech in our 4x weekly newsletter, virtual events with industry experts, and digital guides.
On Gupshup’s services
Gupshup’s hefty client list includes Netflix, Sony, Unilever, Hyatt, Abbott, Citibank, and Khan Academy, with industry verticals spanning real estate, banking, e-commerce, advertising, media, and healthcare.
“The way it works is to think of it as like three layers of the stack: The basic layer is just basic APIs to send and receive messages. But then the intelligence layer on top provides all the workflows, the journeys, the AI—we’ve leveraged LLM, and so on. We’ve launched some of our own capabilities as well, so that when a user sends a message, you can reply to it. So, that entire intelligence layer is unique and key. And then the third layer is—we specialize by industry verticals.”
Gupshup, Sheth explained, is powered by a variety of LLMs, like Meta’s Llama 2, OpenAI’s GPT-3.5 Turbo, Flan T-5, and more. While Gupshup’s ACE LLM uses some of these foundation models, it’s also “meticulously adapted for specific industries and functions.” The company’s domain-specific models, he says, are designed to keep industry and function-use cases in mind.
“It’s the next level up. We customize it by industry, we fine-tune it, and we train them, let’s say for banking, for commerce, for retail, and so on,” he said.
On the evolution of messaging apps
“Because everybody’s using messaging—there are two options: either you kind of force them to use the old-style browser and web, which wasn’t designed for this thing, or you use the app that actually is popular on the small devices, on small screens—and that’s the messaging app.”