![\"Retail](\"https:\/\/dmsretail.com\/RetailNews\/wp-content\/uploads\/2022\/05\/RETAIL-ONLINE-TRAINING-728-X-90.png\")
<\/a><\/p>\n<\/p>\n
The promise of AI to deliver smarter, faster, cheaper business outcomes moves from theory to reality, reports Miya Knights, Retail Technology Publisher<\/strong><\/p>\n Even though <\/span>ChatGPT<\/span> launched over two years ago, it was evident at a US retail technology trade show held last week in New York that retailers are still coming to terms with its impact.<\/span><\/p>\n The artificial intelligence (AI) powered chatbot captured imaginations with its advanced natural language processing capacity, which allows it to understand and generate human-like text.<\/span><\/p>\n So, Generative AI (GenAI) was the main buzzword at <\/span>last year\u2019s Retail\u2019s Big Show<\/span>, hosted by the US <\/span>National Retail federation<\/span> (NRF). But the tech was even more prominent this year.<\/span><\/p>\n John Furner, <\/span>Walmart US<\/span> President and Chief executive and NRF Chairman, declared in the 2025 conference keynote: \u201cIt feels like that’s all we’ve been talking about the last few years.\u201d<\/span><\/p>\n Furner led a discussion with Azita Martin, Retail & Consumer Packaged Goods (CPG) Vice President and General Manager for <\/span>NVIDIA<\/span>\u00a0[pictured], allowing the chipmaker to take centre stage.<\/span><\/p>\n NVIDIA’s graphics processing unit (GPU) systems are optimised for parallel computing, which is essential for training deep learning models, such as ChatGPT\u2019s large language model (LLM).<\/span><\/p>\n \u201cAI requires a very different kind of compute,\u201d Martin explained. \u201cOur GPUs powered the breakthroughs of many companies, including OpenAI, Perplexity and Meta. But we’re more than GPUs.\u201d<\/span><\/p>\n She wanted to highlight the accelerated software component of NVIDIA’s computing platform, enabling developers to build and deploy AI applications at scale through its cloud partners.\u00a0<\/span><\/p>\n Martin introduced the AI Blueprint for retail shopping assistants, using NVIDIA AI Enterprise and Omniverse\u2122 platforms to create sophisticated, automonous AI-powered digital agents.\u00a0<\/span><\/p>\n Utilising NVIDIA NeMo\u2122 microservices, these agents can handle multimodal customer queries with intelligent, context-aware responses. They also integrate with LLMs for natural interactions, and offer state-of-the-art simulations for GenAI and 3D product visualisation.<\/span><\/p>\n She explained how the shift from basic machine learning (ML) programming to the deep learning transformer architecture enabled the training of language models on massive amounts of data. Taking advantage of GPU compute further laid the groundwork for GenAI developments.<\/span><\/p>\n \u201cSo today, companies are fine tuning these large language models with their specific domain specific data, the documents, videos, data and are creating basically digital knowledge,\u201d she continued. \u201cThese AI agents are there to help make your employees more productive.\u00a0<\/span><\/p>\n \u201cA shopping assistant is a good example. Another would be having AI agents writing the responses for your customer service people. They’re still there to check it to make sure that it’s correct. But it’s making your customer service agents much more productive.\u201d<\/span><\/p>\n Martin also cited examples of companies, including <\/span>L’Or\u00e9al<\/span>, Walmart, and <\/span>Lowe’s<\/span> deploying these technologies today. L’Oreal uses digital twins and GenAI models, such as Stable Diffusion, to create ads and marketing campaigns faster.<\/span><\/p>\n Walmart uses NVIDIA’s data science acceleration libraries to improve forecast accuracy, and Lowe’s has created digital twins of its 1,700 stores to optimise layouts and improve sales and revenue, according to simulated shopper footfall and behaviour.<\/span><\/p>\n NVIDIA\u2019s second annual \u201cState of AI in Retail and CPG\u201d <\/span>survey<\/span> found over 80% of retail and CPG companies were either using GenAI or piloting projects. Content generation in marketing and advertising, as well as customer analysis and analytics were among the top use cases.<\/span><\/p>\n A further 87% said AI had a positive impact on increasing annual revenue, while 94% said AI has helped reduce annual operational costs. But less than half (45%) were using AI to reduce supply chain costs, even though 97% said spending on AI would increase in the next fiscal year.<\/span><\/p>\n As an online shopping, voice assistant and cloud infrastructure development leader, <\/span>Amazon<\/span> outlined what AI will deliver for its customers, merchants and business this year.\u00a0<\/span><\/p>\n Doug Herrington, Chief Executive of Amazon Stores Worldwide, said: \u201cAI is becoming transformative for business. We haven’t had a technology revolution as large as this since the start of the internet. It\u2019s bigger than the shift to mobile and the advent of social commerce.\u201d<\/span><\/p>\n Citing GenAI customer review summaries, size and fit recommendations and its Rufus shopping assistant, he added: \u201cWe\u2019re up to half a billion questions we wouldn\u2019t have been able to answer before.\u201d<\/span><\/p>\n Beyond customer-facing applications, Amazon is also offering its marketplace merchants the ability to generate product descriptions, imagery and ads using GenAI tools.\u00a0\u00a0\u00a0<\/span><\/p>\n \u201cIt\u2019s also improving our robotics for how they see and grasp things […] and it’s helping manage traffic with our big fleets of mobile robots,\u201d Herrington added.\u00a0<\/span><\/p>\n Google Cloud also discussed its AI roadmap development, highlighting 160 AI models of its Vertex AI platform and its retail relevance for computer vision and agentic use cases.<\/span><\/p>\n While Google\u2019s ChatGPT rival is good for some use case, Google\u2019s Paul Tepfenhart, Global Director of Retail Industry Strategy & Solutions at <\/span>Google Cloud<\/span> told <\/span>Retail Technology<\/span><\/em>: \u201cWith Vertex, it’s not just Google models we’re using, it’s the whole portfolio.<\/span><\/p>\n \u201cBig models are big time expensive, right?\u201d he continued.\u201dLet\u2019s say I need a smaller model that’s specifically trained on all of our HR benefits. Now it’s very cost effective to run, and it doesn’t take a ton of compute. So, we now have moved from massive LLMs to an agentic approach.\u201d<\/span><\/p>\n Tepfenhart envisages agents that are finely tuned for different purposes. \u201cThen agents can call agents, so it can be layered. That\u2019s why you’re going to see massive adoption, because it becomes much more affordable, delivering much better ROI [return on investment],\u201d he added.<\/span><\/p>\n Wayfair<\/span> has partnered with Google Cloud to use its advanced Gemini models on Vertex AI. The home retailer is automating product categorisation to streamline the curation of new and existing listings, using precise image analysis for accurate product attribution.<\/span><\/p>\n The retailer anticipates significant improvements in operational efficiency, projecting a 67% reduction in time-to-market for product listings. The automation process is expected to save hundreds of thousands of dollars annually by eliminating manual tagging tasks.\u00a0<\/span><\/p>\n It also expects, enhanced listing quality and improved customer satisfaction are likely to boost conversion rates by approximately 2%.<\/span><\/p>\n “At Wayfair, we believe GenAI is the key to unlocking the next generation of retail experiences,” said Fiona Tan, Chief Technology Officer at Wayfair, in a statement.\u00a0<\/span><\/p>\n “With Google Cloud, we’ve been able to efficiently scale and enrich our product catalogues, enabling us to support a more seamless and engaging shopping experience for our customers.”<\/span><\/p>\n<\/div>\nHarnessing potential intelligence<\/strong><\/h3>\n
Developing retail blueprints<\/strong><\/h3>\n
State of adoption<\/strong><\/h3>\n
Next-generation innovation<\/strong><\/h3>\n
The era of agentic AI<\/strong><\/h3>\n
Taking AI to market<\/strong><\/h3>\n