As retailers continue to think about the ways artificial intelligence can support greater personalization for their customers, one start-up has stopped to consider how to personalize the AI integration experience for the companies that rely on its technology.
XGen AI has officially launched its Model Marketplace, which will allow retailers to compose customized search and recommendation capabilities with AI systems, the New York-based start-up told Sourcing Journal.
The company, which came out of stealth earlier this year, works to help brands like Furla, Valentino, Rachel Comey and Berle create AI solutions that work well with their business goals, particularly around product discoverability.
You May Also Like
Frank Faricy, the company’s founder and chief executive officer, said he believes traditional software as a service (SaaS) models don’t serve their customers with the precision and accuracy that his company does in the ever-growing AI landscape.
That’s because XGen AI’s verticalized, composable AI approach allows its clients to use the functions they need and ditch the others.
“The Model Marketplace was designed to really change permanently the way [e-commerce teams] think, to give them tools that they’ve never been able to access before,” he said.
The two main products included in the release are XSearch and XRecommend, which each have smaller microservices underneath them to help e-commerce leaders build their own solutions based on their business cases. Faricy said the release includes 34 models, about 60 percent of which fall under XRecommend and the rest of which are associated with XSearch.
XSearch, he said, will aim to help brands and retailers return search results that better fit the mold of what consumers are on the hunt for when they visit a site — whether that’s occasion-based outfits, a specific colorway for a shoe or otherwise.
“XSearch is a toolkit that allows you to leverage many different models in different contexts to accomplish and accommodate different search methodologies. So you can build conventional based e-commerce search, you can build generative search and you can add — like Lego blocks — different tools to ultimately build a search engine that’s right for you in a very short amount of time,” Faricy told Sourcing Journal.
Meanwhile, XRecommend, he explained, aids retailers in ranking the top few products a consumer may be interested in, based on a real-time feed of information that continuously interacts with the AI systems.
Each model has been purposefully paired with automation to ensure e-commerce teams can implement the technology with minimal hassle. Faricy said he knows that e-commerce teams don’t have the technical expertise expected of an AI team, so XGen has chosen to automate the model parameters so the systems can run with minimal input.
But, in the case that an organization has an internal AI team at its disposal, the engineers can help tune the model to the team’s needs even further.
“Everything about [the models] is written for e-comm teams; even the data requirements are written as e-comm teams understand them,” he said. “If there’s an internal AI team, they can actually switch off the automation and have access to everything they would see in a in a core deep tech system to train a model.”
He expects that the majority of the brands and retailers the company works with will use the two toolkits in tandem with one another in an effort to streamline recommendation rankings following customer searches.
A third piece of the puzzle, which does not have a patchwork collection of models underneath it in the same way XSearch and XRecommend do, is XGenerate, which Faricy described as a way brands can “enrich [their] product data coming into the ecosystem with a bunch more product information, [like] micro categories, macro categories, attributes [and] colors in a generative way.”
The company has plans to add other models to its marketplace, which Faricy said could number more than 100 within the next few months. He said he believes the technology will prove a solution to some of the personalization and discovery problems companies face on a daily basis, while also shifting the power into the hands of e-commerce teams and lightening the implementation burden for technology teams.
For him, it seems that the marriage between automation and AI could result in e-commerce leaders becoming “heroes” within their organizations.
SAP Emarsys data shows that 35 percent of consumers indicated they want AI to aid them in discovering new products. And as organizations race against one another, scrambling to retain customers’ attention and testing new personalization use cases every day, Faricy said he knows the demand will persist as XGen continues to build.
“When the iPhone was released, it was like, ‘Are you kidding me?’ [There were] a lot of skeptics. That’s an extreme example, but the difference here is, there’s already an AI wave, where the demand is there. There wasn’t demand for an iPhone [because] no one knew what an iPhone was,” Faricy said. “But there is demand for AI. All we’re doing is codifying that and giving the wave an outlet to actually be used in [this] vertical.”