The Next Holy Grail for Unified Commerce: The AI-Infused Universal Cart

Several years ago, I wrote an article that was published in RIS News about the universal cart being the holy grail for unified experiences in retail. RIS News, which was a great publication in its day, has passed, as has the initial wave of expectations for a universal cart. With AI, we now look to a turbocharged, highly personalized version of what to expect from a universal cart.

What is a universal cart?

When I think of “universal cart,” I envision an independent component living in the cloud that can serve as the shopping basket and checkout element of transactions for ALL consumer-serving applications implemented within a retailer.

The goal: a consistent shopping experience for the consumer, delivering the retailer’s branded experience across all points of engagement. Along with a more unified shopping experience, this brings consistency in elements such as pricing, applied promotions, access to inventory, product information, and personalized experiences for the consumer across all selling channels. This certainly sounds easy, and some of these elements have been delivered. But others remain elusive. Also, with the availability of new AI tools, how the consumer shops is quickly changing, impacting what the universal cart can become.

Recent achievements in the quest for a true universal cart

Over the past few years, as an industry we have delivered on some of the promises of a universal cart, but to varying degrees, with most retailers and software vendors focusing on specific areas that play to their strengths. Today, you will see various elements of a unified cart, from access to pricing across all sales channels to unified promotions to access to all sellable inventory across the retailer’s enterprise, as well as shared views of important customer data such as purchase history and consistent views of product-related data such as product descriptions, ratings and reviews, and suggested items.

Today, we see a few fragmented solutions that are sometimes referred to as a universal cart and other times as use cases representing unified commerce solutions. Typically, these solutions are piecing together separate applications via highly customized integration or require that all selling components are from a single vendor. The ability to integrate disparate systems to provide a universal cart is an initial step on the journey to a stand-alone, autonomous universal cart.

Yet consumers are beginning to change how they shop, with more focus on leveraging AI tools such as ChatGPT to help identify a specific product or a range of products based on their interests. As has historically been the case, retailers are trying to keep up with changing consumer demand and looking to adapt to ensure their existing customers remain loyal to them. Ultimately, we are all going to have to adapt to this new shopping paradigm – and a new universal cart approach where the consumer is the real owner of their own cart. 

AI as the supercharger for moving beyond the universal cart

AI is the catalyst for new shopping patterns, and it is transforming how retailers interact with customers and manage their inventory. A key driver of this change will be the customer’s cart, which will provide greater insight into what the customer wants and create new options for how to convert items in the cart into sales.

AI will force retailers to gain cleaner access to data from all elements of their business. This data will then be the driver in AI-based decisions that will optimize inventory, pricing, and buying decisions. It will also be the driver for more personalized shopping experiences and help guide retailer interactions with customers – based on what the customer has placed in their universal cart.

Along with access to data, retailers need better access to a shared cart across all points of engagement with the consumer – including being able to share data across a variety of software vendors. AI has already helped in this area, including at Aptos, where we are now delivering our next generation of integration options, which are accelerating the implementation of a common integration layer that can be shared across Aptos and non-Aptos applications.

Consumers also need access to a single, unified cart – independent of any selling source – that any application can access, update, and check out from.

AI will play an important role in personalizing this experience in numerous ways, including by automating the content and timing of personalized communications based on what is in the consumer’s cart and how long it’s been there. This could include automated notifications about sales, price increases, or low inventory related to items in a consumer’s cart. AI will help ensure updated content is timely delivered to customers based on the contents of their cart to help drive sales conversion.

The AI technology that will make this a reality

AI is already transforming how businesses operate, and it will continue to evolve at warp speed. Yet one thing remains consistent: The efficacy of large language model AI systems is dependent on the data the LLMs can access. The larger the quantity of clean, accessible data available to AI, the better the models perform.

Thanks to Aptos’ investments in cloud data warehouses, our Aptos ONE unified commerce platform has a significant advantage in feeding highly desirable data into a segmented AI ecosystem. Fine-tuning pretrained models (created by large AI research labs) with comparatively static item data will allow Aptos ONE to teach the underlying models what keys and terms are relevant to the retailer’s specific line of business. Attributes like stretch and size should have a bigger impact for an apparel retailer than a jeweler. Using traditional data imports to fine-tune existing models results in greater relevancy.

Next-generation API specifications like Model Context Protocol, Agent2Agent, and Universal Commerce Protocol allow Aptos to adopt hyperresponsive RAG workflows at the cutting edge of agentic AI operations.

However, the cloud platform is not the only story. The offline data architecture of the Aptos ONE Store Selling Assistant on store device hardware will unlock fully local and self-contained use cases for AI-driven search and AI-augmented suggested selling, maximizing retailer hardware investments while ensuring the highest degree of data privacy.

The power of what the universal cart can become could significantly shift how consumers shop, as well as create a new playing field that retailers will need to adapt to in order to maintain customer loyalty and win the sale. The AI-infused universal cart will impact us all!