Rugs Direct says shoppers who use its on‑site conversational AI, Ask Cleo, add items to cart at rates 8–9x higher than other users — a clear signal that conversational search can surface more qualified intent, even when overall engagement remains limited.
The assistant, developed with Constructor and rolled out by parent company Elevated Brands Group, debuted on lighting brand Lightopia in May and on Rugs Direct in early June. Sharon Gautschi, chief merchandising officer at Elevated Brands Group, reported the add‑to‑cart lift but stressed that most visitors still use classic search; the company is focused on increasing engagement with the assistant.
How Ask Cleo works and how it was built
Ask Cleo is an agentic AI implemented in two on‑site roles. The AI Shopping Assistant (ASA) provides a higher‑funnel, site‑level conversational search experience; the Product Insights Agent (PIA) lives on product detail pages to answer specific, item‑level questions.
The assistant accepts natural‑language descriptions, asks clarifying questions and supplies practical guidance. On Rugs Direct that can mean explaining materials, thickness and dimensions; on Lightopia it ranges to technical lighting questions such as dimmability, ceiling height guidance and bulb compatibility — areas Gautschi described as “much more technical” than rugs.
Elevated Brands Group trained Ask Cleo using its owned content: product specifications, articles and blog posts used for PR and SEO. Gautschi estimated the initial training phase took about four to six weeks. She emphasized data integrity — consistent, accurate specs across the assortment — as a key enabler of reliable responses, and said the company also trained the agent on brand voice so it behaves like a “well‑informed friend with great design, style and taste.” The team continuously tests responses and returns optimized examples to their third‑party partner when adjustments are required.
Early signals, limits and channel context
The most eye‑catching metric is the 8–9x add‑to‑cart rate among Ask Cleo users. Gautschi also said more engaged conversations — shoppers who ask multiple questions — correlate with stronger KPIs: higher add‑to‑cart and purchase rates.
Those lifts stem from a smaller user base, however. Gautschi cautioned that classic search still drives the majority of sessions, so the company is prioritizing tactics to increase adoption of conversational flows.
Off‑site AI platforms such as ChatGPT, Anthropic’s Claude and Perplexity are beginning to refer traffic, but Gautschi estimated AI referrals currently represent under 10% of visits to Rugs Direct and Lightopia. She expects that, if AI platforms deliver more qualified referrals, ROI and ROAS from those channels could outpace traditional paid programs.
For context, Rugs Direct ranks No. 622 in Digital Commerce 360’s Top 1000 Database and No. 492 in Digital Commerce 360’s AI Rankings.
What this rollout means for marketers and merchants
The Rugs Direct rollout highlights a practical implementation path: pair structured product data and owned content with a third‑party agentic AI to deliver both site‑level and PDP‑level conversational experiences. The reported four‑to‑six week training cadence shows that, with organized data, initial deployment can be relatively fast.
Operationally, two factors determine quality: data integrity and an active testing loop. Conversational agents reflect the quality of catalog and content data, so keeping specs current and instrumenting interactions to capture attribution and conversion metrics are basic requirements. Equally important is defining the assistant’s decision rules and tone to match brand expectations.
Immediate actions for e‑commerce teams: prioritize a catalog audit, align owned content for FAQ‑style responses, set performance tracking for conversational flows, and run experiments that nudge users from classic search into guided conversations. Measure not just engagement but downstream conversion and average order value to judge true lift.
Watch two signals as this work matures: whether conversational search adoption grows beyond an early minority of sessions, and whether off‑site AI referral traffic becomes a consistently qualified, high‑ROI source compared with traditional paid channels. Those outcomes will determine if agentic on‑site assistants move from an experimental tool to a core channel for discovery and conversion.
