The adoption of Artificial Intelligence (AI) among Indian enterprises and startups is on the rise, and many such organisations are in the early stages of their AI-based transformations.
At TechSparks 2023 in Delhi, AI sector experts came together to document this trend and uncover how Indian organisations are making sense of AI for business use cases.
Sangeeta Bavi, Executive Director, Digital Natives, Microsoft set the agenda for the discussion by describing the three basic ways startups are riding the AI wave.
“The first way to leverage AI is to build large foundational models and multi-modal scenarios yourself. The second way is for B2B startups to provide AI-based solutions to customers to boost their revenues,” she said.
“The third way is for startups to use AI to optimise their own operational efficiencies to improve customer support, improve sales conversions, and other use cases.”
The panellists then discussed how startups should decide whether to opt for an off-the-shelf AI model or build a custom one, and arrived at a consensus that the stage of the startup plays a critical role in determining the answer.
“You can’t always use off-the-shelf models. However, at an early-stage startup without significant funding or a developed in-house data science team, it’s good to explore if an external SaaS player can give you the AI capabilities you need to build your Proof of Concept (POC). This can help you understand if it is indeed possible to solve the problem statement you’re going after,” explained Alok Kumar, Head of Engineering, Perfios.
According to him, early-stage startups would find it difficult to beat global industry standards of AI models and should try building their own models only if they feel that nothing on the market is good enough.
“Businesses should first explore how existing models can solve their problem, find the delta, and then figure out if this delta can be addressed by building in-house,” he added.
.thumbnailWrapper
width:6.62rem !important;
.alsoReadTitleImage
min-width: 81px !important;
min-height: 81px !important;
.alsoReadMainTitleText
font-size: 14px !important;
line-height: 20px !important;
.alsoReadHeadText
font-size: 24px !important;
line-height: 20px !important;
AI for ecommerce and fintech
The panel also explored the various advantages and efficiencies AI can bring to business models, especially for interacting with users.
“Initially, some startups built their own chatbots from scratch. Now, with ChatGPT, you don’t need your own Natural Language Processing (NLP) setup to understand the user. As long as you understand conversation flow and have ChatGPT built into it, your bot can understand the user,” said Gautam Rajesh Shelley, Founder, AiSensy.
In fact, through WhatsApp chats, an ecommerce company can suggest new items to users based on what they’ve bought before, he explained.
A similar approach can be taken in the world of fintech, where conversational AI can help users raise issues through speech in several regional languages.
“All customers have to do is talk, and the bot will understand their problems, create a ticket, and figure out the best way to solve it. This happens in real-time and across various regional languages—a setup that works well for all our merchant partners,” said Ritesh Mohan Srivastava, Chief Data Scientist, BharatPe.
Edited by Kanishk Singh
