Before ChatGPT revolutionised AI conversations, one of the largest home improvement retailers in the world, Lowe’s, had already teamed up with OpenAI to enhance its AI capabilities.
“We were partnering with them on our AI efforts even before ChatGPT,” shared Amit Kapur, VP of AI and data analytics at Lowe’s. “When it launched, our collaboration deepened, focusing on enhancing product data quality.”
The collaboration addressed a critical challenge – ensuring accurate product data for over 16 million weekly customer transactions. Accurate product descriptions not only reduce errors but also enhance inventory planning and customer satisfaction. Lowe’s used OpenAI’s GPT-3.5 model for prompt engineering and fine-tuning, resulting in improved search accuracy and a 60% increase in error detection.
Nishant Gupta, senior director of data, analytics, and computational intelligence, said, “Excitement in the team was palpable when we saw results from fine-tuning GPT-3.5 on our product data, and we knew we had a winner on our hands!”
Even numbers back this success. By fine-tuning GPT-3.5, Lowe’s achieved a 20% boost in accuracy and streamlined operations by reducing associates’ workloads for error vetting. These advancements have directly impacted Lowe’s e-commerce growth, making product searches more reliable and customer experiences smoother.
Early Adopter of GenAI in Retail
In the competitive landscape, Lowe’s stands out not just for its early adoption of generative AI but for its strategic use of partnerships and in-house innovation. While companies like Wayfair focus on multimodal models for product matching and recommendations, Lowe’s has built a robust AI ecosystem by using NVIDIA’s computer vision tools and Google’s analytics.
Fiona Tan, CTO at Wayfair, while talking about the evolving AI landscape, stated, “Generative AI enables personalised shopping journeys for each customer every time they click.” Yet, Lowe’s has set itself apart by seamlessly integrating OpenAI’s models into its operational backbone.
Competitors like IKEA have introduced AI tools such as IKEA Kreativ for 3D space visualisation, and Target and Walmart have adapted AI for better inventory and customer experiences.
However, Lowe’s internal developments, such as omnichannel order management systems and self-checkout terminals, underline its commitment towards scalable and reusable solutions.
“We conceptualise reusable components so our solutions evolve with business and technology needs,” Kapur further said.
Bengaluru: The Backbone of Lowe’s AI Evolution
Lowe’s Bengaluru team has been a critical driver of its AI journey and has played a critical role in developing and deploying cutting-edge solutions.
“Many of our core systems, including the omnichannel order management and self-checkout terminals, were built from the ground up by our engineers in Bengaluru. These solutions have provided us with unmatched scalability and flexibility,” Kapur pointed out.
The Bengaluru team’s efforts include designing the cart and checkout engines for Lowe’s stores, integrating AI-powered tools, and creating microservices-based architectures. “We build components like a Lego block framework…allowing us to replace or upgrade functionalities without disrupting the entire system.”
Moreover, the team has been instrumental in implementing AI-driven product tagging solutions fine-tuned with OpenAI models. “The innovations from our Bengaluru team have not only improved efficiency but also enhanced the shopping experience for millions of customers per week,” Kapur noted.
This Indian hub also leads AI advancements in areas like inventory optimisation and predictive analytics, which in turn shapes Lowe’s strategy of becoming an evolving omnichannel retailer. The team in Bengaluru has prioritised reusability, scalability, and customer-centric innovation to allow Lowe’s to stay ahead in the AI game.
Looking ahead, Lowe’s plans to upgrade newer OpenAI models by focusing on the flagged errors to ensure continuous improvements in data accuracy. The retailer’s next steps include using AI for productivity initiatives and expanding its in-house AI frameworks.
