Google Cloud’s Gen AI Innovations for Retail

News
January 11, 2024

Google Cloud Unveils New Generation AI Solutions for Retailers

Google Cloud has recently unveiled its latest generation of AI products specifically designed for retailers. With the aim of personalizing online shopping experiences and streamlining back-office operations, these new gen AI tools are set to revolutionize the retail industry. From Conversational Commerce Solution, which allows retailers to embed gen AI-powered agents on their websites and mobile apps, to the Catalog and Content Enrichment toolset that automatically generates product descriptions and metadata, Google Cloud is making significant strides in enhancing the retail experience. In this blog post, we will explore the features and benefits of these new gen AI products and discuss the potential impact they may have on the retail industry.

Google is aiming to revolutionize retail with a touch of generative AI. At the National Retail Federation’s annual conference in New York City, Google Cloud revealed its latest gen AI innovations. These new products are crafted to enhance online shopping experiences with personalized touches and to optimize back-office processes for retailers, marking a significant step in retail technology advancement.

Regarding their performance as advertised, it remains uncertain — this writer didn’t have the chance to test the new tools before their debut this morning. (Their launch is scheduled for sometime in Q1.) Nonetheless, the cascade of announcements clearly highlights Google’s vigorous pursuit to captivate the gen AI customer base.

Google Cloud’s latest offering, the Conversational Commerce Solution, enables retailers to integrate gen AI-driven agents into their websites and mobile applications, akin to a bespoke version of ChatGPT for each brand. These sophisticated agents engage with shoppers through natural language dialogue, offering personalized product recommendations that align with individual customer preferences.

While branded chatbots are not a novel concept, Google emphasizes that their ‘sophisticated’ models, such as PaLM, are the driving force behind these agents. These advanced agents are capable of being fine-tuned and tailored using retailers’ proprietary data, including product catalogs and website content, enhancing their functionality and customization potential.

Enhancing the Conversational Commerce Solution, Google Cloud introduces its innovative Catalog and Content Enrichment toolset. This suite employs gen AI models, including the aforementioned PaLM and Imagen, to automatically create detailed product descriptions, metadata, and categorization recommendations from just a single product image. Additionally, this toolset allows retailers to generate fresh product images from existing ones or to create AI-generated visual representations of products based on their descriptions, further streamlining and enriching the retail experience.

When eBay introduced a comparable AI-powered feature that converts product images into descriptions a few months ago, it quickly faced criticism from sellers. They raised concerns about its performance, highlighting issues with misleading, overly repetitive, and in some instances, blatantly inaccurate descriptions.

Amy Eschliman, Google Cloud’s Managing Director of Retail, was questioned about the steps Google has implemented to address inaccuracies in AI-generated content. Although she did not cite specific initiatives, she emphasized Google’s ongoing commitment to refining its tools. Eschliman also underscored the crucial role of human supervision in the workflows of the Catalog and Content Enrichment toolset.

It’s certainly reassuring to hear about the human review, especially in high-stakes situations. After all, it’s not far-fetched to imagine that an AI-generated image or description in a product catalog that’s misleading could put a retailer in a difficult position with customers, or even lead to accusations of false advertising.

Eschliman explained, ‘Human-in-the-loop is a crucial best practice for enterprise use cases. It ensures high quality, mitigates risks associated with bias, and fosters trust and transparency. Additionally, it aids in model improvement and ongoing training, while adhering to regulatory and business policies.’

In a concurrent announcement, Google unveiled a retail-specific version of its Distributed Cloud Edge device. This managed, self-contained hardware kit aims to ‘reduce IT costs and resource investments’ in retail gen AI applications. While Distributed Cloud Edge has been a long-standing service, its focus is now more directly on retail applications. The edge cluster, available in configurations ranging from single-server to multi-server, is designed to integrate seamlessly into various retail environments — from convenience stores and gas stations to fast casual eateries and supermarkets — to support customers’ gen AI applications.

Eschliman highlighted the robust capabilities of Google Distributed Cloud Edge, noting, ‘With its control plane operating locally, this system ensures uninterrupted operations for retailers, even during brief internet outages that can last days. Retailers are now equipped with a compact cluster of Google Cloud-managed nodes, adaptable for installation in almost any store setting. This fully managed hardware and software solution enables retailers to seamlessly run their existing software alongside distributed AI, ensuring mission-critical operations are maintained in-store at all times.

Google has announced that details on pricing and availability will be made available in Q1.

Post the pre-briefing, a pertinent question arises: Are retailers truly eager for gen AI technologies? It appears so, particularly among the retail giants.

For instance, Walmart recently revealed its substantial investment in gen AI search capabilities to enhance query context understanding and enable shoppers to search by specific use cases, like ‘unicorn-themed toddler birthday party’. Amazon, on the other hand, has been utilizing gen AI to summarize customer reviews, assist sellers in crafting product descriptions and image captions, and improve the efficiency of finding clothes that match a customer’s size.

In a survey conducted by Google, the company reports that 81% of retail decision makers perceive a strong ‘urgency’ to integrate gen AI into their operations. Furthermore, 72% believe they are prepared to implement gen AI technologies immediately. This readiness is particularly evident in areas such as customer service automation, marketing support, product description generation, creative assistance, conversational commerce, and enhancing store associate knowledge and support.

However, given the recent turbulent deployments of gen AI in the retail sector (for instance, Amazon’s review summaries that overstated negative feedback), I remain skeptical about the retail industry’s widespread and rapid adoption of gen AI — whether from Google Cloud or other providers. It seems we’ll just have to wait and observe how this unfolds.

Conclusion

Google Cloud’s new gen AI products for retailers offer exciting possibilities for enhancing the online shopping experience and optimizing back-office operations. With features like Conversational Commerce Solution and the Catalog and Content Enrichment toolset, retailers can personalize interactions with customers and automate various tasks, ultimately improving efficiency and customer satisfaction.

While the effectiveness of these tools remains to be seen, Google’s commitment to continuous improvement and human review processes instills confidence in their reliability. As the retail industry continues to evolve, embracing gen AI technology may become crucial for staying competitive and meeting customer expectations. With Google Cloud leading the way, retailers have the opportunity to leverage the power of AI to transform their businesses and deliver exceptional experiences to their customers.

How do you currently generate product descriptions, metadata, and categorization suggestions? Would you be interested in automating these processes using gen AI models? How do you see gen AI shaping the future of the retail industry? Do you believe it will become a crucial component for retailers to stay competitive and meet customer demands? Share your insights below!