A sofa is a measured decision — will it fit the room, survive the dog, make it up the stairwell. HOLLOU answers those questions from your own listings and delivery pages, captures the shopper who isn't ready to buy tonight, and says so plainly when your catalog can't answer.
Furniture shoppers research across days and devices before they commit. The questions that stall them are specific — and a search box or generic chatbot can't take them, because the answers live in dimension tables, variant lists and delivery pages, not keywords.
Search returns products, not answers. The shopper is left doing arithmetic with a tape measure and a product photo — and doubt abandons a $2,000 cart, not price.
Filters can't hold “a performance fabric that survives a dog and won't fade in a west-facing window.” So shoppers open eight tabs, lose the thread, and leave.
White-glove or curbside, stairs, assembly, packaging removal — it's on your delivery page, and almost nobody reads it. The doubt rides along to checkout, and often wins.
HOLLOU is catalog-grounded: every dimension, fabric, price and policy in an answer is retrieved from your own site and product data. When it runs out, the agent says so.
The agent reads assembled and boxed dimensions from your listings and works them against the shopper's room, doorway or stairwell. When a clearance is tight, it says so and asks the shopper to confirm the measurement before ordering — a returned sectional costs more than a cautious answer.
Fabric grades, finishes, leg options, lead times — every combination you publish is ground truth. The shopper says what they care about; the agent narrows to the versions that match, and never invents one.
White-glove versus curbside, coverage by area, assembly, packaging removal — answered from your delivery and policy pages at the moment the doubt appears, not three clicks away.
Furniture shoppers rarely buy on the first visit. When someone needs a swatch, a custom configuration or a human eye, the agent captures the request — room size, fabric constraints, budget — and hands your team a warm lead instead of an anonymous bounce.
When the shopper is ready, the agent adds to cart and hands them to your own checkout. It never touches payment.
A modeled example — not customer results. The store, products and prices below are invented. The real agent answers from your catalog.
“Will the Marlow 3-seat sectional fit a living room that's 12 by 14 feet?”
The listing puts the Marlow at 112″ wide and 65″ deep with the chaise. On the 14ft wall that leaves roughly 28″ of walkway past the chaise — workable, but snug if the door opens inward. Worth confirming: the chaise is left-facing only on this model.
“I have a dog. Which of the Marlow fabrics will actually survive?”
Two of its fourteen fabrics carry the catalog's heavy-use rating and are listed as stain-resistant.
“Do you deliver to 75035, and can you take the old sofa away?”
White-glove delivery covers that ZIP at $149 — placement and packaging removal included, per the store's delivery page. Old-item removal isn't mentioned on the site, so I can't confirm it. Want me to send that question to the store with your details?
“Could I get the Copeland dining table in a 78-inch length?”
The catalog lists the Copeland at 64″ and 72″ only — there's no 78″ option published, and I won't guess at custom work. If you leave your email, I'll pass this to the store as a custom-length quote request.
Notice the last two answers. Where the catalog runs out, the agent stops — and turns the question into a lead instead of a guess.
One script tag, alongside your existing site — theme, product pages, cart and checkout stay as they are. Furniture is one of four verticals we've modeled in depth; the others are on the use-cases page, and pilot pricing is published openly on plans.
From your own listings. The agent reads the assembled and boxed dimensions you publish and reasons against the shopper's stated room or doorway. If a measurement isn't in your catalog, it says so rather than estimating one.
Variant-heavy catalogs are the point. The agent treats each combination you publish as fact — fabric grades, finishes, leg options, lead times — and narrows by what the shopper actually cares about. No dropdown maze.
It says it doesn't know, and offers to capture the question as a lead or quote request. Your console flags those gaps, so unanswered questions become a to-do list for your catalog content.
No. It qualifies and books. Shoppers who need a swatch, a custom configuration or design help get captured with their context and handed to your team — the agent does the midnight shift, your people do the judgment.
A 30–60 day pilot on a controlled slice of your traffic, with success criteria agreed in writing before launch and no long-term contract. If it doesn't show what we hoped, it ends there.
Send us your store URL. We'll index a slice of your furniture catalog and show you the agent answering real questions about your actual products — before you commit to anything.
No catalog credentials needed for the first preview