When Purva Gupta emigrated to the U.S. from India and did some online shopping here, she found herself surprisingly frustrated. Why was it so hard to seek out the precise form of dress she was searching for? Considering it is likely to be an immigrant problem resulting from a language or cultural barrier, she did some hefty local research to “test her hypothesis.” But after speaking with greater than 1,000 women, she confirmed there was definitely a disconnect between how consumers experienced and perceived products, and the way brands and retailers described such products. This compelled her to create an AI-driven solution that bridges the merchant-consumer gap.

WWD caught up with Gupta to debate Lily AI, which she cofounded with chief technology officer Sowmiya Chocka Narayanan. Lily AI recently received a $25 million Series B investment, which it’s using to expand into mid-market retail e-commerce brands across home, beauty and fashion.

WWD: How did your prior background lead you to cofounding Lily AI?

Purva Gupta: While I’m an economist by education, my experiences working in India at Saatchi & Saatchi after which at a start-up paved the best way for what became Lily AI. While at Saatchi & Saatchi, I used to be inspired by the ability of the emotional connection between a client and a brand. I spotted I desired to work in technology and alter people’s lives, and concluded that if I desired to create my very own company, then the core of the issue it was trying to resolve needed to be an issue I experienced personally and deeply connected with.

WWD: On the subject of online search engines like google and yahoo, how big is the gap between merchant speak and consumer speak?

P.G.: From a language perspective, the words real people use are much more colloquial and infrequently more nuanced than the usual words utilized by retailers and types to explain their products. Consumers have unique emotional contexts and perspectives. After they detail what they’re on the lookout for, they use a wealthy, personalized vocabulary that features dimensions like trends, occasions and styles. Ultimately, it boils all the way down to product details, which, in merchant-speak, specifically refers back to the product attributes that exist in a retailer’s product taxonomy.

For instance, a brand may tag what a consumer calls a “summer wedge” as a “supple leather upper resort wedge sandal,” a “back-supporting mattress” as a “perfect sleeper ultra-plush hybrid gel mattress” or a “lightweight summer foundation” as “Stay-in-Place Flawless Wear Cashmere Matte Foundation.” The examples are countless but show how consumers and retailers approach language otherwise.

WWD: How do you train Lily AI’s algorithms to be more consistent with how consumers search?

P.G.: Humans are all the time within the loop. Our domain experts have backgrounds in retail (merchandisers, stylists, marketers), and this team stays on top of the newest trends to tell probably the most robust consumer-friendly product taxonomy leading to improved trend discovery. They’re always conducting in-depth research into micro- and macro-trends, textile and color trends and social media trends. Armed with this information, they train machine learning to make sure the Lily AI product taxonomy is built to match consumer trends with relevant attributes.

Data scientists and AI engineering experts are also an element of the numerous humans behind Lily AI’s unique “consumer-oriented” product taxonomies versus pure-play automation, always refining models and ensuring the very best in data quality, accuracy and relevance. This mix of experts is always training and refining the algorithms, leading to an ML that matures over time, constantly getting ‘smarter’ and ever more accurate with every training input.

WWD: What are the outcomes of this?

P.G.: We’ve got compiled a proprietary library of over 20,000 consumer-oriented words spanning attributes, synonyms and trends, and we use this ever-expanding, vast data asset to tell our product taxonomy. Doing so, we are able to keep pace with the evolving voice of the buyer. Amongst a few of our brands, which we should not at liberty to reveal, now we have seen a 3.5 to 9 percent increase in online order conversion, a 2 to five percent increase in product detail page views and a 3 to 10 percent increase in demand.

WWD: With GenAI basically, users are learning the worth of an authority prompt. Do you think that this growing expertise will help improve online shopping searches?

P.G.: One shouldn’t have to be a prompt engineer to seek out what they need. The nice news here is that search engine technologies and platforms will proceed to evolve so that customers don’t need engineering degrees to buy online. We’re within the early innings of GenerativeAI, and as now we have already seen within the one 12 months since ChatGPT launched and altered our world, it would only recuperate, not to say, safer.

But even with great prompts, for search to “find” what an individual seeks, we still need the relevant product details and attributes to be properly labeled to power the invention.

WWD: How does Lily AI help with demand forecasting and what have been some tangible client advantages?

P.G.: Planning and forecasting are prioritized focus areas for lots of our clients as a consequence of the huge margin increases to be realized from improved pre-season and in-season models. At Lily AI, our demand attributes help retailers to boost product design, improve replenishment and allocation models and deliver an assortment that maximizes margin opportunity.

One among our multibrand clients projected $7 million to $48 million in top-line revenue increase from leveraging the Lily AI-improved product attribution data of their forecasting models. One other global retailer estimated a possible to cut back weighted average percentage error, or WAPE, by 20 percent and improve gross margin by $300 million across all brands.

WWD: How is AI evolving and the way can brands and retailers harness it to maximum effect?

P.G.: AI for retail just isn’t latest. Be it data-driven analytics, applying machine learning in inventory planning, supply chain all of the approach to powering customer experiences through recommendations, chatbots and detecting anomalies in retail security, machine learning has played a task in retail for quite a while.

The underlying technology has been evolving rapidly and getting smarter by the day, and the wave of deep learning excites us for its ability to learn to make connections between input and output and requires less spoon-feeding than earlier ML techniques needed. And now generative AI has pushed ML capabilities from analyzing or classifying existing data to having the ability to create something entirely latest, including text, images, audio, synthetic data and more.

That said, so as to effectively harness the worth of today’s powerful suite of AI, it will be important to all the time start from deeply understanding the use case and the issue we try to resolve, having the proper, accurate data, after which the skillsets of the team and infrastructure to find a way to experiment to reach at the proper solution.

At Lily AI, we perform 1000’s of experiments before we push the outperforming models into production. Our platform can be built with the pliability to swap in/out the proper models for the issue in hand. Our vision is to bring humanity to shopping and we’re excited to proceed to innovate and draw on our retail AI expertise to assist global brands and retailers thrive.