May 2026 Summaries

Visual search is revolutionizing ecommerce product discovery by allowing shoppers to find products based on their visual appearance rather than text descriptions alone. This technology integrates both text and images into a unified model, enhancing conversion rates by providing more relevant search results and recommendations. Traditional text-only search systems fail to capture the nuances of product appearance, such as style and texture, leading to mismatched results. In contrast, multimodal search, which processes text and images in the same mathematical space, offers a more accurate understanding of products, thereby improving search relevance and driving revenue. Companies like Marqo have developed multimodal AI systems specifically for ecommerce, resulting in significant revenue increases for retailers who adopt this technology. For example, brands like Fashion Nova and Redbubble have seen substantial improvements in search conversion rates and revenue by deploying multimodal search. This approach is particularly beneficial in visually driven categories like fashion and home goods, and it also enhances product recommendations by understanding and visually pairing products effectively.

Coveo and Marqo are two distinct platforms serving enterprise ecommerce teams, each with unique approaches to product discovery. Coveo, established in 2005 as an enterprise search company, is built on a keyword-based engine enhanced with machine learning for re-ranking and personalization, and has expanded into multiple verticals. In contrast, Marqo, founded by former Amazon engineers, focuses exclusively on ecommerce with an AI-native architecture designed to understand shopper intent through product-native intelligence, processing text, images, and attributes within a unified model. Marqo's approach allows for immediate product understanding and relevance without relying on accumulated behavioral data, making it particularly effective for long-tail queries and new products. The platform's capabilities extend to multimodal search and advanced merchandising controls, offering significant revenue uplift as demonstrated by clients like Fashion Nova and Kogan. While Coveo provides a broad, multi-vertical solution with recent additions like generative AI, Marqo's specialized focus on ecommerce discovery aims to deliver faster and more measurable outcomes for retailers.

By 2026, the debate has shifted from traditional keyword search versus semantic search to a more nuanced comparison between AI-layered and AI-native search architectures in ecommerce. While most ecommerce platforms have integrated some form of AI to enhance search capabilities, many still rely on generic models that lack a deep understanding of specific product attributes and shopper behavior, often leading to revenue shortfalls. AI-native search, exemplified by platforms like Marqo, is emerging as a superior alternative due to its purpose-built models that comprehend both textual and visual product attributes, enabling immediate recognition of new products and nuanced shopper intent. This architecture not only improves relevance and reduces zero-result rates but also boosts conversion rates, revenue per session, and average order value by aligning search results with both shopper preferences and business priorities. As retailers pivot towards AI-native solutions, they witness significant gains, especially in handling long-tail, descriptive, and conceptual queries, highlighting the limitations of AI-layered approaches that depend heavily on pre-existing behavioral data.

Ecommerce search is a crucial component of online retail, influencing conversion rates significantly higher than general browsing, yet many retailers still treat it as a secondary feature rather than a primary revenue driver. The guide emphasizes the importance of treating search as a revenue infrastructure, highlighting the substantial financial gains realized by companies such as Fashion Nova, Redbubble, and Mejuri through optimized search systems. Key best practices for 2026 include moving beyond outdated keyword matching to semantic search that understands the intent behind queries, integrating images and text for a seamless user experience, and fine-tuning search models to cater to specific product catalogs. Additionally, effective search merchandising strategies such as implementing boost and bury rules and aligning search with business calendars can enhance product visibility and align with business objectives. The document also underscores the significance of mobile-first design and precise measurement of search metrics to continually improve the search experience. AI-native search platforms, which offer a unified approach to handling search, recommendations, and merchandising, are presented as the future of ecommerce search technology, with the potential to significantly elevate the shopping experience and increase revenue.

AI-native ecommerce search represents a significant evolution in the online retail search landscape, shifting from traditional keyword-based systems to architectures entirely built around AI models designed specifically for product discovery. Unlike AI-layered systems that add AI to existing infrastructures, AI-native search integrates purpose-built models that understand product attributes, categories, and shopper intent, enhancing conversion rates and revenue. This approach resolves common issues with legacy systems, such as handling descriptive queries, overcoming the cold-start problem, and providing a comprehensive understanding of visual and textual product details. Retailers like Redbubble and Fashion Nova have reported substantial revenue increases by adopting AI-native search, which also enables faster deployment and less manual configuration compared to traditional systems. This architecture not only improves search outcomes but also enhances other product discovery experiences like merchandising, category organization, and recommendations, by genuinely understanding the products involved.

Marqo and Bloomreach are two distinctive platforms evaluated by ecommerce teams for product discovery, each offering unique features and capabilities. Bloomreach, established in 2009, has evolved into a comprehensive Commerce Experience Cloud with modules for search, marketing automation, and content management, leveraging its AI layer, Loomi, to enhance these functionalities. However, its search relies on behavioral data which can be limiting for new products or low-traffic queries. In contrast, Marqo is a purpose-built, AI-native product discovery platform focused solely on ecommerce, offering a seamless integration of search, merchandising, and personalization without requiring prior behavioral data. Marqo's models are trained specifically on a retailer's catalog, ensuring immediate and intelligent product ranking from the onset. Additionally, Marqo provides native multimodal search capabilities, enabling integrated text and image queries, which can be advantageous for visually-driven categories. While Bloomreach's extensive platform may suit enterprises seeking a consolidated vendor for multiple functions, Marqo appeals to ecommerce retailers prioritizing rapid, specialized solutions for product discovery, offering documented revenue uplifts and swift implementation timelines.

Kogan.com, one of Australia's largest online retailers, successfully leveraged Marqo's AI-driven search platform to generate $10.1 million in incremental revenue, highlighting the tool's efficacy in managing extensive, complex catalogs. Operating with over 16 million products and experiencing constant catalog churn due to its hybrid model of first-party inventory and third-party marketplace products, Kogan faced significant challenges with traditional search systems, which struggled with vocabulary diversity, long-tail queries, and cross-category intent. Marqo's dedicated AI model, trained specifically on Kogan's catalog, offered instant understanding of new products, cross-category intelligence, and effective handling of long-tail queries, thus eliminating the cold-start problem and improving search relevance across the board. This transition not only led to a substantial increase in revenue but also prompted Kogan to expand its use of Marqo's full product suite, underscoring the platform's capability to provide enterprise-scale solutions for large and diverse retailers.

Resale platforms face significant challenges with traditional search engines due to their reliance on behavioral data, which is ineffective for the fast-paced, unique, and diverse nature of resale inventories. Traditional search engines, designed for long-lasting retail products, struggle as resale items are one-of-a-kind, have no prior behavioral data, and sell quickly, often before search algorithms can adjust. AI-native search solutions like Marqo address these issues by focusing on product attributes, descriptions, and images, enabling instant relevance for new listings and improving product discovery. Marqo's AI-driven approach has shown significant improvements in click-through rates, add-to-cart rates, and revenue across various branded resale platforms by understanding product characteristics rather than relying on accumulated clicks. This method not only benefits resale platforms but also offers a scalable solution for broader e-commerce challenges, where rapid product turnover and personalized discovery are increasingly prevalent, positioning AI-native search as a crucial tool for future-proofing online retail strategies.

Marqo and Cimulate represent two distinct approaches to AI-native ecommerce product discovery, with Marqo offering a platform-independent solution and Cimulate being integrated into Salesforce's ecosystem following its acquisition. Marqo's model is built specifically for ecommerce, leveraging deep product understanding and behavioral data to drive search, merchandising, recommendations, and conversational commerce, while maintaining platform independence and allowing integration with various commerce platforms like Shopify and Adobe Commerce. In contrast, Cimulate, now part of Salesforce's Agentforce Commerce, employs a synthetic data approach for its CommerceGPT model, focusing on text-based discovery and context-aware recommendations but with limited visual search capabilities. Marqo has demonstrated significant revenue impacts for retailers, with documented success across diverse categories, while Cimulate's results remain largely anecdotal and concentrated in fashion and apparel. Marqo's Commerce Superintelligence offers a unified intelligence layer, enabling seamless integration and rapid implementation, whereas retailers considering Cimulate must be prepared to commit to Salesforce Commerce Cloud.

Marqo offers a platform that automatically builds a dedicated AI model for retailers by training it specifically on their product catalogs, eliminating the need for any manual setup or a machine learning team. This model differs from a generic shared search model by providing contextually relevant and specific responses based on the retailer's unique vocabulary, product relationships, and category structure. By addressing the cold-start problem, Marqo's AI understands new products and categories from the outset, without relying on accumulated behavioral data, which is crucial for low-traffic queries and seasonal items. The AI continuously improves with incoming data, requiring no manual retraining, and ensures privacy by using data exclusively for the specific retailer's model. This dedicated AI not only enhances search functionality but also powers recommendations, merchandising strategies, and other commerce touchpoints, offering immediate and tailored results from day one.

As the ecommerce search landscape evolves, the role of AI-native platforms becomes increasingly crucial for enhancing user experience and driving conversions. Unlike traditional keyword-based systems, where AI functions as an overlay, AI-native platforms reimagine the retrieval process by deeply integrating AI to understand shopper intent through both text and visual signals. This innovative approach offers significant improvements in handling vague or intent-driven queries and adapting to dynamic product catalogs without extensive manual tuning. The guide emphasizes the importance of selecting platforms that are purpose-built for ecommerce, highlighting the need for quick implementation and live testing to ensure relevance and performance on specific catalogs. Additionally, the integration of post-purchase intelligence is highlighted as a key factor in maintaining customer loyalty and providing a seamless shopping experience. Successful case studies, such as Fashion Nova and Redbubble, underscore the tangible benefits of adopting AI-native search solutions, which include rapid experimentation, measurable revenue impacts, and enhanced customer engagement.

Marqo and Nosto offer distinct approaches to product discovery for ecommerce retailers, with Marqo focusing on AI-native search and Nosto providing a bundled personalization platform. Nosto, branded as experience.AI, integrates keyword-based search through its acquisition of Searchnode alongside other features like personalized emails and pop-ups, aiming to consolidate multiple solutions into one platform. Marqo, on the other hand, builds its platform from the ground up with AI at its core, offering semantic and multimodal search capabilities that leverage deep product understanding and behavioral data. Marqo's models are trained specifically for each retailer's catalog, providing immediate product understanding and improving search relevance from the start, whereas Nosto relies on accumulating behavioral data to enhance search performance over time. Marqo's AI-driven approach has demonstrated significant revenue uplifts for its clients, with reported successes like a $130M revenue increase for Fashion Nova, highlighting its strength in ecommerce product discovery and personalization.

Algolia has been a popular search API since 2012, but its general-purpose architecture poses limitations for ecommerce teams seeking specialized product discovery to drive revenue. As of 2026, several alternatives offer tailored solutions with varying strengths and limitations. Marqo stands out for mid-market and enterprise retailers with its AI-native platform, providing dedicated AI per retailer and native multimodal search, resulting in significant revenue increases for clients like Fashion Nova and Kogan. Constructor focuses on behavioral ranking and merchandising, while Bloomreach offers a comprehensive Commerce Experience Cloud. Coveo excels in cross-platform search, Klevu provides quick implementation for smaller retailers, Elasticsearch requires engineering investment for custom solutions, Searchspring integrates search with merchandising for mid-market retailers, and Nosto combines search with personalization tools. When evaluating these options, considerations include platform understanding of specific products, activation timelines, and the ability to handle visually driven categories, with Marqo delivering the largest published revenue results.

Commerce Superintelligence represents a transformative approach to AI in retail, offering a system that deeply understands products, shopper behaviors, and personalization to enhance every stage of the shopping journey. Unlike previous generations that relied heavily on keyword search or behavioral ranking, Commerce Superintelligence starts with a product-native intelligence that comprehends the nuances of each item, integrating this with behavioral data to refine results continuously. This system is characterized by six architectural requirements: product-native intelligence, full-journey intelligence continuity, unified cross-modal retrieval, zero-shot product competency, embedded commercial optimization, and visual product reasoning. These requirements ensure that AI can understand and prioritize products accurately from the moment they enter a catalog, without reliance on accumulated shopper interactions, optimizing for both shopper relevance and business value. Marqo exemplifies this approach by providing a platform that integrates these capabilities, resulting in significant improvements in conversion rates and revenue for major retailers. By shifting from a behavioral to a product understanding foundation, Commerce Superintelligence offers a more robust, adaptable, and comprehensive solution for modern ecommerce challenges, enabling advanced search, merchandising, recommendations, and conversational commerce across the entire customer journey.

Constructor and Marqo are e-commerce product discovery platforms that utilize AI to enhance conversion, revenue, and shopper experience through advanced search capabilities. While Constructor emphasizes detailed, rule-based merchandising control and behavioral optimization, Marqo focuses on a unified intent understanding, integrating text, image, and product attributes in a multimodal ranking system for enhanced personalization and ranking optimization. Marqo's proprietary models, trained on each retailer's catalog and actual shopper behavior, have shown significant revenue uplifts, such as Fashion Nova's $130M increase, compared to Constructor's outcomes like Sephora's $40M revenue lift. The comparison also highlights differences in search quality, merchandising strategy, implementation speed, and customer results, with Marqo offering quicker integration and scalable merchandising controls. Both platforms provide conversational product discovery features, but Marqo's architecture supports a more seamless integration of visual and multimodal search, making it particularly effective in visually driven industries.

Ecommerce's reliance on behavioral data for product discovery has led to significant challenges, with the traditional clickstream approach creating an information bottleneck that fails to accommodate new and unique products. This system, which prioritizes past shopper behavior over product attributes, often results in missed opportunities for retailers, such as undiscovered inventory and homogenized curation that undermines brand identity. To address these issues, the industry is shifting towards AI-native product discovery powered by Commerce Superintelligence, which emphasizes understanding the intrinsic attributes of products and using behavioral data to refine insights. This approach enables retailers to handle new products and one-of-a-kind items effectively, ensuring they are visible and relevant from day one. The transition is not merely an upgrade but a structural necessity for modern ecommerce, as demonstrated by companies like Archive, which saw significant improvements in metrics such as add-to-cart rates and revenue per user after implementing this new model. As retailers like Fashion Nova and Mejuri have shown, adopting Commerce Superintelligence can lead to substantial revenue growth and improved conversion rates, highlighting its significance as the next generation of ecommerce measurement.

Marqo and Algolia are two prominent search platforms used in ecommerce, each built on distinct philosophies that cater to different needs. Algolia, established in 2012, is known for its fast, keyword-based search API, and has recently incorporated AI capabilities with its NeuralSearch product, which enhances search relevance by combining keyword matching with neural embeddings. However, it requires a threshold of behavioral data to activate its AI features, which can be limiting for new or low-traffic retailers. In contrast, Marqo is an AI-native platform designed specifically for ecommerce, providing a dedicated AI model for each retailer that understands their unique product vocabulary and purchase patterns from the outset, without needing a behavioral data warm-up. Marqo also excels in multimodal search by processing text and images in a single query, offering significant advantages in visual-driven categories like fashion and home goods. While Algolia serves a wide array of industries, Marqo's focus on ecommerce is evident in its product design, supporting features like merchandising integration, conversational commerce, and rapid implementation that have led to substantial revenue uplifts for its users.

The blog post explores the distinction between AI-native and AI-enhanced search platforms in the context of product discovery, emphasizing the architectural differences and their implications for performance and scalability in retail environments. An AI-native platform is built from the ground up with artificial intelligence as its core, allowing it to handle queries with deep understanding and without keyword constraints, whereas AI-enhanced platforms add AI capabilities to pre-existing legacy systems, often resulting in limitations due to their reliance on traditional keyword-based indexing. The AI-native architecture facilitates "Commerce Superintelligence," a concept where product understanding and shopper behavior data are integrated seamlessly across all commerce interactions, enhancing search quality, conversion rates, and operational efficiency. The post argues that for large retailers, selecting an AI-native platform is crucial for achieving maximum potential in search-driven revenue and adapting to modern consumer behaviors that involve multimodal queries.

Sibbi, a conversational commerce agent developed by Marqo, transforms the online shopping experience by utilizing Commerce Superintelligence to guide shoppers from product discovery through post-purchase interactions. Unlike generic chatbots, Sibbi is specifically trained on individual retailers' catalogs, ensuring it understands product details and inventory in real-time, which allows for accurate recommendations and seamless transactions. Sibbi not only interprets textual and visual search inputs to provide relevant product suggestions but also enhances cross-selling with complementary product recommendations based on genuine product relationships rather than collaborative filtering. It extends its usefulness beyond the purchase phase by managing order status inquiries, returns, and future purchase suggestions, thus maintaining a continuous and intelligent shopper-brand interaction. This innovative approach has already demonstrated significant revenue and conversion rate improvements for major retailers like Fashion Nova and Mejuri, with Sibbi's deployment being rapid and its results measurable within two weeks.