SUMMARY
Inc42 and Google Cloud collaborated to host a roundtable exploring the transformative impact of genAI across various industries in India
The session covered many critical topics, including: the productivity boost that comes with genAI implementation, the impact of genAI on existing technologies, the challenges of genAI adoption
The roundtable brought together startup leaders from varied sectors like fintech, proptech, healthtech, ecommerce and more
The deafening buzz around generative AI (genAI) is gradually taking a concrete shape and turning old assumptions about technology on their head. Now that transformative tech is in our midst, it is the right time to explore how it redefines existing technologies, how the Indian startup ecosystem uses genAI capabilities to drive profit and productivity and how businesses should prepare for a genAI future. To understand the staggering impact of genAI development across businesses and industry segments, Inc42 and Google Cloud organised a roundtable titled “How Generative AI Is Fuelling A Cross-Industry Transformation In India.”
The session covered many critical topics, including:
- The productivity boost that comes with genAI implementation
- The impact of genAI on existing technologies
- The challenges of genAI adoption
Moderated by Rajan Nagina, head of AI practice at Newgen Software, the roundtable brought together startup leaders from varied sectors like fintech, proptech, healthtech, ecommerce and more.
Among the attendees were Kunal Gupta, senior vice-president at magicpin; Nitin Jain, cofounder & chief business officer, OfBusiness; Devnagri cofounder Himanshu Sharma; Fashinza tech head Diwas Sharma; Vipin Singh, head of technology at PropTiger.com & Housing.com; Shakti Goel, chief architect and data scientist at Yatra Online; Animesh Sharma, chief technology officer at Indifi; Gaurav Bagga, SVP (product & engineering) at Pristyn Care; Saurabh Dwivedi, SVP (technology) at MobiKwik; Kapil Malhotra, head of customer engineering at Google Cloud, and Harsh Sarohi, SVP and technology head at TradeIndia.com.
During the discussion, Kunal Gupta of magicpin cited how investing heavily in genAI enhanced productivity and led to quick issue resolution. “We recently had a system outage due to a legacy code issue, but there was no assigned owner. The developer concerned had already left the company. Nevertheless, we recreated the entire functionality within an hour using genAI. We dockerised [created a lightweight environment to run the application and its dependencies] and deployed it to production and resolved the issue quickly,” he explained.
Decision-makers and industry leaders often consider genAI a silver bullet. It also motivates them to explore how best to use its capabilities for a wide range of sure-fire solutions, whether related to cost savings, efficiency building or coping with talent shortages through automation.
MobiKwik’s Saurabh Dwivedi is one of them, deep-diving beyond the obvious. “Our goal is to create more engaging experiences for our customers. For instance, we have launched Lens, a tool for presenting financial data graphically to help users understand it better. Now, we are looking at genAI to summarise this information more effectively.”
Match Capabilities With The Right Use Cases For GenAI Success
GenAI has powerful capabilities and great potential. But its success will depend on how well it caters to the specific use cases pursued by a business. Content or code generation with the help of genAI is super-hyped and appears straightforward. However, syncing data with this new technology requires careful planning to ensure excellent outcomes.
“If you use the RAG model [retrieval-augmented generation is used to optimise the output of an LLM], you won’t get 100% accuracy. Plus, there will be a lot of AI hallucinations or incorrect output based on perceived notions. So, the deciding factor is whether your use case demands 100% accuracy or 70-80% accuracy will be fine. If you are okay with 70-80%, there is hope because you can trim your expectations [from the genAI model], and your purpose is served,” said Shakti Goel of Yatra Online.
Kapil Malhotra of Google Cloud concurred, underscoring the criticality of use case-specific needs.
“It’s not about code generation alone. As startups scale rapidly, they accumulate significant tech debt from applications developed years ago. So, one of the use cases [for genAI adoption] is not just to roll out a feature faster but also to reduce your tech debt.”
“GenAI is not just a utility; it has to be part of your application. Similarly, when we talk about overall adoption, it is not just about the model. None of the models will be completely accurate anytime soon. Much work is pending on your data, which you will continue to do. On the other hand, models will continually evolve; token sizes will increase [to overcome text limits that can be inputted] and new challenges will arise. But it is crucial to prioritise your tech debt and focus on developer productivity,” added Malhotra.
Despite the surmises and the soaring hype, there is visible apprehension regarding the quick adoption of genAI for faster results and fast-mover advantages.
“We are treading cautiously on that,” said Animesh Sharma of Indifi. “It’s not that I have to fit genAI into everything. There’s customer support, for instance. If the mobile app has a comprehensive UI/UX and a category in place, customers can solve their problems, and I don’t have to put a genAI solution there. I am happy with a rule engine wherever it works well. Wherever machine learning (ML) works great, wherever it can lift our business, we are doing it.”
Nevertheless, the genAI wave is helping startups think on their feet and adapt faster. More importantly, genAI is transforming ML as we know it, turbocharging breakthrough innovations.
Nitin Jain from OfBusiness summed it up well, underlining the opportunities that the new technology has ushered in. “Using genAI, we are communicating with our clientele a lot better and a lot more. Instead of getting in touch once every 15 days, we communicate with them daily using a lot of genAI, ML and algorithmic tools. GenAI just got in and made ML adoption a lot better.”
