Can AI Bring Sales Out Of “The Paleolithic Age Of E-Commerce”?

While shopping for party clothes for his daughter, Shridhar Marri had his epiphany about where e-commerce should head.

His daughter had just graduated from Syracuse University, and she had very specific requirements for the parties to follow. She wanted “effortless cool” that was “minimalistic, elegant, and not extravagant.” And none of the beauty products could be tested on animals. There was context and nuance in what she sought, yet websites and marketplaces just provided thousands of products without relevance.

Ah-ha! Marri realized that digital commerce was lagging behind in-store sales in a vital way. While businesses wanted customers to increasingly use digital channels, customers often craved personalized guidance and advice. These two imperatives seemed at cross-purposes. Could they be reconciled?

Marri established his company, Flyfish, to try. He explains what happened next. “We built the platform and created integrations for the enterprise stack, so that any brand could come in, have their data ingested, link their product catalogs, and then deliver a consultative sales experience. Along the way, generative AI emerged and we saw this as a godsend. It can make the experience even more compelling. So we integrated with Large Language Models (LLMs) to create a really consultative sales experience at the time of buying.”

The Role of LLMs vs. Other Forms of AI

While much of current writing on AI focuses on generative platforms like ChatGPT, oftentimes the best solutions mix forms of AI. Flyfish has created its AI stack to take this approach.

Marri lays out the approach. “There’s algorithmic AI at the basis of the stack to make tailored recommendations based on the seller’s catalog as well as structured inputs. Those can include the CRM’s data about the buyer, data about past conversations, and information from the ERP about stock availability and pricing. Then the LLM rides on top to humanize the response. And then on top of that we need to tailor the LLM to the sales use case in a specific industry vertical, so that it has the right language for persuasiveness. We take an open source LLM and build on top of it for that outcome.”

What This Means for E-Commerce’s Future

Marri believes that e-commerce is at a juncture. “It’s become cold and impersonal. While there may be a plethora of choices, it’s cumbersome to find the right product.” He continues, “The fundamental thing is to bring the high touch experience when you’re online. If you walked into a store, proactive sales agents would help you in a totally appropriate way. You don’t rely on that advice all the time. Here, you can have a tagalong advisor which the consumer can sort of believe, especially when they seem objective and knowledgeable. That experience of having a personal advisor can really grow.”

This is a vision shared by Bill Gates. He told a Goldman Sachs event in March, “Whoever wins the personal agent, that’s the big thing, because you will never go to a search site again, you will never go to a productivity site, you’ll never go to Amazon again.”

Whether or not Flyfish helps to create that personal agent, the vision seems to be spreading and shopping experience may change fast. Marri concludes, “In just a few years in the future, we’re going to see 2023 as a primitive time for digital sales. The paleolithic age of e-commerce. It’ll be like looking at the early web and just seeing text and hyperlinks.”

Today, AI tools for sales are typically talked about as sales enablement tools. If Marri or others can accomplish the feat of having AI close sales, not just enabling humans to make them, the party-clothes epiphany will truly be something to celebrate.