Retail / Consumer Packaged Goods

Lowe’s Enhances Customer Experience With Gen AI and Digital Twins

Lowe’s

Objective

Lowe’s Home Improvement, a Fortune 100 company operating over 1,750 stores with 300,000 associates, serves tens of millions of customers weekly across an omnichannel network. The company leverages NVIDIA Omniverse™, NVIDIA DeepStream, and NVIDIA GPUs as foundational elements for its comprehensive AI strategy. The goal is to revolutionize customer shopping experiences, empower store associates, and transform merchandising through physically accurate digital twins, generative AI, and computer vision.

Customer

Lowe’s

Partner

Dell Technologies

Use Case

Simulation / Modeling / Design
Generative AI / LLMs

Key Takeaways

  • 98% neutral-to-positive customer reactions achieved with Mylow, Lowe’s AI-powered home improvement virtual advisor
  • 8,000 AI-assisted engineering code reviews completed weekly, accelerating development timelines and app deployment
  • Store associates equipped with an AI-powered store companion tool for instant product knowledge and customer assistance
  • 30,000+ item 3D asset catalog enhanced by AI-powered 2D-to-3D conversion with Hyper3D.ai’s Rodin, reducing creation costs to under $1 per model
  • Proactive customer assistance enabled through computer vision solutions was deployed across stores for real-time support
  • 1,750+ stores with digital twins updated throughout the day, providing advanced visualization, analysis, and simulation capabilities
  • Virtual planogram testing of hundreds of scenarios digitally, making physical resets simpler, faster, and more cost-effective

Adapting to Market and Seasonal Shifts Remains Costly

The home improvement sector of the retail industry is uniquely complex. Vast product arrays, diverse customer needs spanning product discovery to project completion, and efficiently managing large physical store spaces all contribute to challenging dynamics. Big-box retailers must constantly adapt to changing market conditions while optimizing operational efficiency and enhancing customer experiences.

Lowe’s faced several critical operational challenges across its massive retail network. Many customers struggle with project-based shopping—creating an expertise gap that traditional product discovery tools couldn’t address at scale. When customers ask, “How do I fix a leaky faucet?” they need comprehensive project guidance—not just simple product information.

Another key challenge was empowering store associates with knowledge of over 50,000 in-store items. Managing 100,000-square-foot stores creates operational visibility challenges as well, making it difficult to monitor real-time inventory levels and identify when customers need assistance.

Traditional merchandising and store planning are often inefficient and costly. Updating planograms and layouts—which dictate product placement, category adjacencies, and customer pathways—usually requires spreadsheet work and time-consuming, in-person physical resets. These processes limit the ability to test multiple scenarios quickly, making it difficult to adapt across stores with diverse regional and seasonal demands.

Lowe’s

Bridging the Physical and Digital Retail Experience

Lowe’s embarked on a comprehensive AI transformation, building an agile, AI-first retail operation anchored in three strategic pillars: “how we shop,” “how we sell,” and “how we work.”

The company’s comprehensive initiative leverages NVIDIA hardware and software platforms that includes Dell PowerEdge servers from the edge to the core, creating a unified stack for advanced AI, immersive simulation, and edge intelligence across the business. Alongside a mixed hardware infrastructure—including GPUs for AI training in the data center, workstations for digital twin simulation, and GPUs for computer vision at the store and edge—Lowe’s operational transformation was empowered by NVIDIA Omniverse, a platform for developing OpenUSD-based 3D applications. NVIDIA Omniverse enables Lowe’s to create fully immersive, photorealistic 3D visualizations for its store digital twin platform. Micro-data centers with rugged, short-depth Dell PowerEdge XR Series servers in hundreds of stores ensure low-latency, real-time data processing and inferencing for AI-driven experiences.

“How We Shop”—Customer-Facing Generative AI

Lowe's encoded their century-long home improvement expertise in "Mylow," a comprehensive generative AI solution integrated into the Lowe's app and website. Mylow lets customers ask questions about their projects and receive step-by-step instructions, curated product recommendations, and links to relevant Lowe's YouTube content.

Mylow uses a hybrid approach to power its Mylow solution, combining OpenAI commercial models fine-tuned on Lowe's proprietary data. NVIDIA NeMo Guardrails ensure protected and reliable AI-generated responses. By employing retrieval-augmented generation (RAG), Lowe's can provide faster customer service with accurate, real-time product availability and location details.

“How We Sell”—Associate Empowerment

The “Mylow Companion” app was developed and delivered to store associates, providing real-time access to the same generative AI engine that powers Mylow. Associates can access instant product and project expertise across unfamiliar departments, from electrical to hardware sections.

Lowe’s leverages NVIDIA DeepStream for computer vision capabilities, deploying GPUs at store and edge locations for real-time video analytics. Store cameras automatically notify the nearest associate when customers dwell in aisles without support, enabling proactive assistance.

“How We Work”—Digital Twin Merchandising

Lowe’s built physically accurate digital twins of all 1,750+ stores in its network, creating a unified foundation for innovation at scale. Updated daily—and in some cases in real time—these models reflect the dynamic nature of each store and can be customized almost infinitely to meet business needs. NVIDIA Omniverse enables immersive 3D visualization and advanced simulation on top of this foundation, unlocking new ways to explore, test, and optimize the store experience.

The company also developed the Lowe’s Digital Design Center, a prototype Omniverse-based application powered by Dell and NVIDIA infrastructure, including RTX™ workstations and GPUs for digital twin simulation and rendering. The Digital Design Center enables merchandisers to:

  • Visualize real-time product inventory and analytics in photorealistic 3D environments.
  • Virtually optimize bay layouts using the AI-powered “Intelligent Bay Optimizer” with predictive forecasting.
  • Generate thousands of localized, constrained planograms digitally, replacing cumbersome spreadsheets and physical designs with illustrated reports.

To deliver immersive, up-to-date digital twin experiences across all stores—and to ensure product data is accurate and easily discoverable—Lowe’s created a massive 30,000+ item digital asset catalog through internal development and vendor partnerships with companies like Whirlpool and LG.

By integrating the NVIDIA USD Search API directly into its workflow, and with Hyper3D.ai’s Rodin, Lowe’s streamlined asset discovery and enabled AI-powered 3D model generation. This process transforms 2D product images into precise, high-quality 3D models within minutes—at a cost of less than $1 per model.

Cloud streaming deployment is helping Lowe’s realize even more efficiency gains. NVIDIA Omniverse technologies deploy via cloud streaming, enabling users to access the solution via web browsers and eliminating the need for powerful workstations across all user locations.

“At the center of the AI revolution is going to be our people and our teams. I would say that having the right teams collaborating with the right partners like NVIDIA, and making the magic happen is where all it starts.”

Seemantini Godbole
Chief Digital and Information Officer, Lowe’s

Lowe’s

An Extensible Framework for Retail Innovation

Lowe’s is advancing its digital transformation with a built-in-house digital twin platform that unifies spatial, product, and operational data. To power visualization and large-scale applications, the company uses NVIDIA Omniverse and the OpenUSD framework.

OpenUSD serves as the common standard, enabling interoperability across immersive environments and compatibility with a wide range of Omniverse extensions and third-party tools such as Autodesk AutoCAD, Maya, Revit, and Trimble SketchUp. This approach allows Lowe’s to flexibly integrate and update new capabilities as business needs evolve.

This standardized platform approach has given Lowe’s a strong foundation to take advantage of the revolution in AI and generative AI. By unifying data into a single extensible environment, Lowe’s can layer in interactive editing, analytics overlays, and emerging AI-driven capabilities—turning what was once disconnected into a cohesive, scalable system.

Synthetic data generation within digital twins is a core advantage, providing large datasets to safely test AI models, validate workflows, and more. This systematic framework not only supports current operations, but also gives Lowe’s a platform to expand automation and digital initiatives into more areas of retail and logistics.

Optimizing for Today, Building for Tomorrow

Lowe’s stands out as a retail pioneer, demonstrating how a comprehensive AI strategy and robust infrastructure can transform every facet of a business. From customer engagement and associate productivity to complex merchandising and store operations, Lowe’s puts AI-powered solutions to work in each one of its stores, every day.

By embracing digital twins, generative AI, and computer vision, Lowe’s optimizes current performance while actively shaping the future of agile, intelligent retail. This transformation showcases how traditional retailers can leverage cutting-edge AI technologies to create competitive advantages, improve operational efficiency, and deliver superior customer experiences across massive retail networks.

Ready to Get Started?

To learn more about NVIDIA solutions for retail and CPG, contact us.

Related Customer Stories