Runa AI (askruna.ai) has built something genuinely useful for Shopify fashion brands. Its AI Merchandiser scans a brand's product catalog and automatically generates "Complete the Look" bundles and outfit combinations, eliminating the manual labour that a merchandising team would otherwise spend creating styled pairs for every product. Its AI Trend Spotter watches social media to surface viral bundle opportunities. Its AI Personal Stylist chat provides 24/7 sizing and style advice. For a brand that wants to automate the operational burden of visual merchandising without a significant investment, Runa AI delivers measurable relief.
But there is a distinction that the Runa AI feature set obscures, and it is the distinction that determines long-term conversion and retention performance. Runa AI completes the look from the catalog. Elara completes the look for the shopper. Those are not the same operation, and the difference between them is the entire gap between good ecommerce and great retail.
What Runa AI does well
Runa AI's catalog scanning and auto-bundling capability is its core strength. The AI Merchandiser ingests a brand's product catalog, builds compatibility logic between items, and generates outfit pairings without requiring manual product tagging. When inventory changes, the system automatically replaces out-of-stock items with in-stock alternatives, keeping every bundle current without human intervention. The AI Trend Spotter layer adds a social intelligence dimension — monitoring trending looks and automatically building bundles that reflect what is performing virally, which gives brands a faster path from social trend to on-site merchandising than any human team could achieve.
The AI Personal Stylist chat layer provides a conversational interface over this catalog intelligence. Shoppers can ask about sizing and get guidance. They can request outfit suggestions and see catalog-level recommendations. The chat is powered by ChatGPT and draws from the brand's product data.
One Runa AI customer captured the brand's aspiration precisely: "I aimed to replicate the exceptional sales experience of Chanel's flagship store in Paris." The ambition is exactly right. The question is whether a catalog-scanning bundling engine with a ChatGPT-powered chat layer gets close enough to that experience to matter.
The problem with catalog-level personalization
Here is what happens when every shopper sees the same outfit bundle. A shopper with a minimalist wardrobe and a maximalist sensibility see the same "Complete the Look" recommendation for the same jacket. A shopper who is dressing for a beach wedding and a shopper who needs something for a corporate board meeting see the same recommendation. A shopper on her eighth visit to the store and a shopper who arrived for the first time thirty seconds ago see the same recommendation, because the bundle was built from the catalog, not from any knowledge of who is viewing it.
This is the structural limitation of catalog-level outfit completion, and it is the same limitation regardless of how sophisticated the catalog scanning is. The bundle is on-brand. It is not personal. And the experience of fashion at its best — in a physical store with a skilled sales associate, or with an online AI that has actually learned who you are — is profoundly personal. It starts with listening, not with showing.
The Runa AI chat improves on this by allowing shoppers to ask questions. But a chat that is powered by general-purpose LLM technology over a brand's product catalog is doing something categorically different from a styling AI that has built a persistent model of an individual shopper's taste over multiple sessions. The first answers questions from the catalog. The second answers from a deepening understanding of the person asking.
What the Style Graph makes possible that Runa AI cannot replicate
Elara's Style Graph is the architectural difference that changes what "personalisation" actually means on a fashion storefront. Every session a shopper has with Elara — every brief she describes, every look she engages with, every piece she adds to her wardrobe, every occasion she mentions — updates a persistent model of her individual taste. The model compounds. A shopper who has visited a brand's Elara-powered store four times has four sessions of taste signal accumulated in her Style Graph. Her next visit produces recommendations that reflect all four.
This is not behavioral tracking in the conventional sense. Behavioral tracking records what a shopper clicked and infers category interest. The Style Graph records what a shopper actually said she needs, what she responded to aesthetically, and what occasions she is dressing for. It builds an understanding of her style, not just her browsing pattern. Over time, it becomes the equivalent of the personal stylist who already knows her wardrobe, remembers that she hates anything with a stiff collar, knows she has a wedding coming up in October, and has been watching for the perfect occasion-appropriate dress to arrive in the catalog on her behalf.
Runa AI's Trend Spotter knows what is trending on social media. Elara's Style Graph knows what a specific shopper's wardrobe is missing and what she would love next. Both are forms of intelligence. Only one of them is personal.
Integration: live in fifteen minutes, no App Store required
Runa AI is available through the Shopify App Store. This is a convenient discovery mechanism, and the one-click install flow lowers the activation barrier for brands exploring AI styling for the first time.
Elara deploys directly — four lines of code added to the Shopify theme, live in fifteen minutes. The integration is independent of the Shopify App Store, which means no waiting period for App Store review, no feature limitations imposed by App Store guidelines, and no dependency on Shopify's approval cycle when Elara ships product updates. The brand connects its catalog through API credentials in the Elara merchant portal, and automatic webhook-based inventory sync begins immediately.
The fifteen-minute deployment is a meaningful operational difference for brands that want to move fast. It is also a signal about how each platform is built. An App Store listing operates within constraints Shopify defines. A direct integration operates within constraints the brand and Elara define together — including the full configuration of the widget's visual identity, copy, placement, merchandising rules, and occasion priorities through the merchant portal.
Revenue transparency: bundles sold vs. revenue actually generated
Runa AI's analytics show bundle performance — how many "Complete the Look" recommendations were shown, how many resulted in clicks, and how bundle pages perform compared to standard product pages. These metrics tell a brand that the tool is generating activity. They do not isolate the causal revenue increment — the revenue that exists because Runa AI was deployed, measured against the revenue that would have existed without it.
Elara's holdout-based lift report does exactly that. A defined percentage of shoppers are assigned to a control group and browse without Elara for fourteen days. At the end of the period, Elara compares AOV, conversion rate, and total revenue between the two groups. The revenue difference is the lift. Brands receive this report automatically — not as a pitch document but as the actual measurement of what happened in their specific store with their specific customers.
The question of causality matters enormously when evaluating an AI tool's ROI. If conversion improved during the same period the tool was deployed, was that Elara, or was it the seasonal traffic spike, or the email campaign, or the new product launch? The holdout design eliminates those confounders. Both groups experienced the same seasonal conditions, the same campaigns, the same new arrivals. The only thing that differed was Elara's presence. The difference in outcomes is therefore attributable to Elara, not to anything else.
What "Chanel flagship store in Paris" actually requires
The Runa AI customer who wanted to replicate a Chanel flagship experience was describing something specific. The Chanel in-store experience is personal. The sales associate knows you, or she builds a rapid understanding of you, and she makes recommendations that feel specifically yours. You do not receive the same recommendation as the woman who walked in before you. You receive a recommendation informed by the conversation you are having with her, right now, building on whatever she already knows about your style.
Catalog bundling — however intelligent the bundling logic — is not that. A consistently on-brand "Complete the Look" module on every product page is a better version of a static lookbook. It is not a personal recommendation. The gap between those two things is the gap between a fashion ecommerce conversion rate of 1 to 2% and the 23 to 30% that physical retail achieves with personal service.
Elara closes that gap. The Style Graph, the conversational brief, the occasion context, the virtual try-on within the conversation, and the fifteen-minute integration are each part of the same answer to the same question: what would it actually take to give every online shopper the experience that makes physical retail compelling?
Book a demo to see what a compounding personal stylist experience — built on your catalog, live in fifteen minutes — does to conversion, AOV, and the customer relationships that drive retention.