Your website shows the products. It never talks anyone into buying one.

An AI sales assistant for websites does what a good salesperson on a shop floor does — it helps a visitor narrow down the right choice, explains why one option fits them, answers the "is it worth it" objection, and suggests the thing that goes with it. Maya does that on your website by voice or text, reading your live catalogue so every recommendation is real. It's the selling layer your product pages can't be, on the site you already have. (Available on the Booking plan and up.)

Last updated 2026-07-16

Where a product catalogue stops

A catalogue lists. It shows the visitor everything and helps them decide nothing. A shopper who half-knows what they want — "something for a wedding, not too formal, under €200" — has to translate that into filters, click through twenty results, and guess which one actually suits them. Most don't. They get overwhelmed, second-guess, and leave with the tab open on a page they never bought from. The catalogue did its job; nobody did the selling.

Maya begins where the catalogue stops. A product page is a pamphlet on a rack — it tells you what exists and waits for you to work it out. Maya is the salesperson who walks the floor with you: it asks what the occasion is, narrows the options, tells you why this one fits what you said, and handles the reason you were about to put it back. Once the assistant can refine a choice in conversation and explain the fit, "chatbot" stops being the right word.

How Maya sells on your website

It refines the choice in conversation — constraints preserved as they stack — the shopper says "a jacket for a wedding, nothing black, under €200," then "actually, with a bit of stretch" — and Maya keeps every constraint as they add them, narrowing to the few that truly fit instead of dumping a filtered grid. It's the difference between a search box and a person who remembers what you just said.

It explains why an option fits — grounded in what the visitor told it — not "this is popular," but "this one suits what you asked for — it's under your budget, it's not black, and the stretch means it'll move with you on the day." The visitor gets the reasoning a good assistant gives, so they buy with confidence instead of leaving to "think about it."

It answers the objection and offers the right next step — the shopper about to close the tab over price, fit, or shipping gets an informed answer and a reason to continue — the way a floor assistant would talk them through it, not a FAQ link they have to go hunt for. The tab that was about to close stays open.

It suggests the thing that goes with it — the upsell, in context — "you've got the dress for the first event — want me to find shoes that work with it for the second?" Maya carries the context forward and offers the natural add-on, the way a person on the floor would, so the basket grows because it made sense, not because a banner shouted.

It sells from your live catalogue — real stock, real price, never a guess — every recommendation is checked against your actual catalogue on Shopify, WooCommerce, WordPress, Wix, or BigCommerce — so Maya never sells a shopper on something that's out of stock or quotes a price you don't charge. The pitch is always real.

It sells in the visitor's own language, by voice or text — the shopper can talk to it or type, and Maya sells in their language — and follows if they switch mid-conversation without asking them to start over. When your site already has that language, it switches the storefront to the correct version. The foreign shopper who'd have bounced gets the full sales conversation in their own words.

It drives the visitor to the product hands-free — once Maya has narrowed the choice, it takes the shopper there — scrolling and navigating to the exact product page for them, hands-free, no mouse or typing — right up until they choose to add to cart or pay. The recommendation and the route to it are one continuous motion.

It captures the shopper who leaves anyway — not every conversation ends in a sale today. The shopper who wasn't ready leaves a name and what they were looking at, so a near-miss becomes a lead you can follow up — instead of an anonymous exit you never see.

What happens when Maya doesn't know the answer?

The fear with an AI that sells is that it oversells — inventing a spec or promising a delivery date to close the sale. Maya doesn't. It checks every claim against your approved catalogue and business information, and when a shopper asks something it can't verify — whether a fabric is waterproof, whether it'll arrive by Friday — it says so instead of guessing, offers the right next step, and flags the question for you. A sale built on a made-up promise is a refund and a bad review; Maya is built not to make that trade.

What that means for you: your assistant sells hard on what's true and stops cold at what isn't. So it grows the basket without quietly setting up returns, chargebacks, and one-star reviews — which is what a salesperson who'll say anything to close actually costs you. Honest selling is the only kind that compounds.