Lily AI

What is Lily AI?

Lily AI is a product intelligence platform for retailers and brands that turns thin product data into detailed, machine-readable attributes. Its core engine, Lily Max, scans a catalogue for gaps, generates enriched titles, attributes and descriptions, and then tests those changes in controlled experiments before scaling the winners. The goal is better performance wherever shoppers discover products, from Google Shopping and Meta ads to onsite search and AI assistants such as ChatGPT and Gemini.

Purva Gupta and Sowmiya Chocka Narayanan founded Lily AI in 2015, according to FashionUnited. The company began with product attribution for fashion retail, using computer vision and natural language processing, and has since moved into agentic commerce. CMSWire reported in March 2024 that Lily AI had raised $62 million in total, including a $20 million Series B-1 led by Conductive Ventures. Its site names brands such as Coach, M&S, J.Crew, HOKA, Foot Locker and Shiseido, and the platform was active in October 2026.

Key Features

• Lily Max engine: Lily Max runs a continuous loop that detects catalogue gaps, generates enrichment, tests it and scales what works. Lily AI describes it as one shared product model that powers every channel it supports.
• Product attribute enrichment: Lily AI enriches titles, attributes and descriptions using computer vision and language models built for retail. It bridges how merchants describe items and how shoppers actually search, which improves matching across fashion, beauty and home catalogues.
• Google Ads optimisation: Lily AI improves product feeds for Google Shopping, Performance Max, Demand Gen and Search campaigns. The company reports a 28% revenue lift on Google Shopping for one large apparel brand, which is its own case figure.
• Meta Ads optimisation: Lily AI also enriches catalogues that feed Meta Advantage+, retargeting and prospecting campaigns. It cites a 21.4% ROAS lift for a national footwear retailer, a vendor result you should test against your own spend.
• AI discovery and agentic commerce: Lily AI prepares product data so answer engines and shopping agents such as ChatGPT and Gemini can read and recommend items. It also handles SEO and AEO metadata, including titles, alt text and structured markup, according to EcommerceGuide.
• Controlled A/B testing: Lily AI proves lift with A/B and difference-in-differences tests against holdout groups before rolling changes out. That separates the effect of better data from seasonality or changes in ad spend.
• Onsite search and recommendations: Enriched attributes feed onsite search, filters and recommendations, not just paid channels. Lily AI reports a 28.3% onsite revenue lift for one luxury brand after enriching its product attributes.
• Gap analysis and performance insights: Gap Analysis scores a catalogue for missing or weak data, while AI Performance Insights tracks revenue from AI-driven discovery. Merchandisers can then prioritise the product groups where better data should pay off fastest.

Pros

• Proven with large retailers: Lily AI lists brands including Coach, M&S, J.Crew, HOKA and Shiseido, and press reports name Macy’s and Bloomingdale’s. That record suggests it handles the scale and complexity of enterprise catalogues.
• Results backed by testing: Lily AI measures lift with A/B and difference-in-differences tests rather than visibility scores alone. Marketing teams get revenue and ROAS evidence they can take to finance before scaling a change.
• No replatforming required: Lily AI works with existing product feeds, PIMs, commerce platforms and agencies. It needs only your current titles, attributes and descriptions, and the company says most teams launch within days.
• Covers paid, onsite and AI channels: One enriched catalogue serves Google Ads, Meta Ads, onsite search and AI answer engines. That avoids maintaining separate data fixes for each channel and lets improvements compound across them.
• Clear data commitments: Lily AI states that data is encrypted in transit and at rest, access is least-privilege, and catalogue or ad data is never sold. Customers keep ownership of their ad accounts, since Lily operates inside them.

Cons

• No public pricing: Lily AI prices by catalogue size, channels and ad spend, and asks you to book a demo for a quote. Smaller brands cannot judge affordability without a sales conversation, though one directory reports a $200 starting price.
• Retail-only focus: Lily AI is built for retail and brand catalogues in sectors such as apparel, footwear, beauty and home. Businesses selling services or non-catalogue products will find little use for it.
• Depends on source data quality: Results rely on the quality of existing product images and descriptions, as FitGap points out. Poor photos or sparse source data limit how much the models can infer and enrich.
• Needs human review: AI-generated attributes may still need checking for brand voice and accuracy, according to FitGap. Teams should plan reviewer time, especially for regulated categories such as skincare claims.
• No guaranteed AI placement: Lily AI itself says no tool can guarantee products appear in ChatGPT or other AI assistants. It improves data readiness, but visibility on those surfaces stays outside any vendor’s control.

Who is Using Lily AI?

• Retailers and brands: Enrich large catalogues across fashion, beauty and home.
• Performance marketers: Improve Google Shopping and Meta campaign results.
• Feed and merchandising managers: Fix product data gaps and onsite search.
• Agencies: Run catalogue-aware retail media for several clients.

Pricing

Lily AI uses custom pricing based on catalogue size, chosen channels and ad spend, and does not publish list prices (official FAQs, checked October 2026).

FitGap lists sterling tiers that appear to describe a different product, so we have ignored them. Lily AI’s FAQs say engagements usually start with a scoped pilot, so ask for the pilot scope, success metrics and contract length in writing before you sign.

What Makes Lily AI Unique?

Lily AI treats every piece of product enrichment as a hypothesis to test rather than content to publish. Lily Max writes the attributes, runs holdout experiments across Google, Meta, onsite search and AI discovery, then scales only what lifts revenue. Feed tools move data and content generators write copy, but few combine retail-trained attribution with built-in proof of impact across paid, organic and agentic channels.

Compatibilities and Integrations

• Ad platforms: Google Ads (Shopping, Performance Max, Demand Gen, Search) and Meta Ads (Advantage+, catalogue, prospecting).
• AI surfaces: Prepares data for answer engines and shopping agents such as ChatGPT and Gemini, plus Google AI Overviews.
• Existing stack: Works alongside current product feeds, PIMs, commerce platforms and agencies without replatforming.
• Onsite: Feeds enriched attributes into onsite search, filters and recommendations.

Lily AI Tutorials

Most teams move from first call to a live pilot in a few days.

1. Book a demo or request the free feed audit on lily.ai.
2. Connect your product feed, ad accounts and storefront data.
3. Review the gap analysis to see which products lack key attributes.
4. Launch a scoped pilot on chosen channels with holdout testing.
5. Scale the enrichments that show measured lift across the catalogue.

How We Rated It

Summary

Lily AI suits mid-size and enterprise retailers whose product data limits performance on Google, Meta, onsite search and AI assistants. Its main limitation is opaque pricing, which makes it hard for smaller brands to judge value early. The testing-first approach is its strongest argument, because it ties enrichment to revenue rather than visibility scores. Start with the free 500-product feed audit on a category with high ad spend, so any lift shows up quickly.