How Generative AI is Revolutionizing Data-Driven Decisions
The advent of Generative AI (GenAI) is transforming industries by enhancing the way organizations and their customers interact with data. Unlike traditional AI, GenAI offers a new paradigm for innovation in data-driven decision-making, according to experts at a recent panel discussion at the YourStory Tech Leaders’ Conclave in Bengaluru.
The panel, featuring Ashwin Sekar of InCred Financial Services, Ramesan Kumar Katreddi of Jumbotail, and Kedar Gupte of Myntra, and moderated by Rajesh Ramdas of Databricks, covered various insights into the strategic use of GenAI.
At Jumbotail, a B2B marketplace connecting sellers and kiranas, GenAI’s multi-modality has significantly enhanced communication abilities, especially for those not well-versed with technology. “This includes vision and voice capabilities that cater to users communicating in vernacular languages, thus improving our click-to-rate improvements through Generative AI,” Katreddi revealed.
In a different sector, InCred’s Ashwin Sekar highlighted how GenAI aids in employee onboarding for their rapidly growing team. Beyond standard applications, InCred leverages GenAI to combat fraud, crafting synthetic datasets to bolster their machine learning models’ fraud detection capabilities.
Myntra is applying GenAI to personalize search experiences on their e-commerce platform. Still, Kedar Gupte stressed that the most beneficial applications lie in developer productivity and query optimization. “This not only covers query plan assessments but also enables more users to optimally leverage our data platform,” he noted.
One prevalent challenge highlighted was the trust in large language models (LLM), a concern addressed by ensuring rigorous data quality validation. “Creating clear data lineage and definitions helps improve LLMs’ outputs,” said Sekar, underlining the necessity of maintaining data integrity.
Highlighting the essential role of data democratization, Gupte pointed out the architectural considerations crucial for GenAI adoption. “Understanding the data’s information graph plays a pivotal role, enabling LLMs to efficiently analyze and extract insights from vast datasets,” he explained.
Despite the progress, the path to widespread LLM deployment in production remains exploratory. Jumbotail’s Katreddi shared insights into their approach to verifying LLM correctness, emphasizing continuous process improvements. Gupte paralleled this sentiment in the context of e-commerce, where the alignment of data quality, model training, and pipeline efficiency directly influences the success of AI-driven recommendations.
In conclusion, the session underscored the importance of adopting good engineering principles and fostering a culture of rapid iteration to effectively leverage GenAI technologies. As organizations continue experimenting within various domains, the consensus is clear: Generative AI represents a strategic lever for elevating data-driven decisions, predictive analytics, and operational efficiencies across industries.
