Question 1

What is AI site search for ecommerce and how is it different from keyword search?

Accepted Answer

AI site search reads a shopper's intent in plain language and returns ranked, explained products. It handles synonyms, budget, size and follow-up questions. Keyword search just matches words, so it struggles with long-tail queries. Sellarix handles both natural questions and exact SKUs.

Question 2

How does conversational search reduce zero-result pages?

Accepted Answer

Sellarix understands what a query means, so dead-end wording still finds relevant in-stock products. Zero-result rescue rewrites failing searches and autocomplete heads off typos before the shopper hits enter, turning a silent revenue leak into a path to purchase.

Question 3

Does Sellarix support visual and image search for online stores?

Accepted Answer

Yes. Shoppers search by photo and Sellarix combines visual similarity with colour, material, price and stock, so matches are products you actually sell. Visual search hands off to the same recommendation and checkout tools, going from camera to purchase in one flow.

Question 4

Does AI size and fit guidance reduce returns and improve conversion?

Accepted Answer

Size and fit guidance is built to cut wrong-size returns, while conversational discovery gets shoppers to the right product faster. Virtual try-on and fit tech are being adopted across retail to lift conversion and curb returns (CNBC, 2026). We instrument everything so you can measure it on your store.

Question 5

How much does Sellarix AI search and recommendations cost per month?

Accepted Answer

Pricing is published and starts at £99 per month for conversational search, scaling to the full platform for mid-market and multi-store brands. There's a clear path for SMBs, not a demo-only enterprise funnel, and advanced AI search isn't gated behind a top tier.

Question 6

How is Sellarix different from Algolia, Constructor, Bloomreach or Rezolve?

Accepted Answer

Sellarix unifies search, recommendations, visual search, fit, bundling, guided selling and CRO on one data spine, so signals compound. It's also model-agnostic and agent-ready. We publish honest, sourced benchmarks rather than asserting unverifiable accuracy or uplift claims.

Question 7

Does Sellarix work with Shopify and small or mid-market stores, or enterprise only?

Accepted Answer

Both. Sellarix connects to Shopify, WooCommerce, Magento, BigCommerce, PrestaShop or a CSV, with published pricing from £99 per month. There's a clear path for SMB and mid-market, so a single Shopify store and a multi-store brand can both start and grow.

Question 8

Does Sellarix learn from my own catalogue and customers or from vendor averages?

Accepted Answer

Sellarix learns on your data. Ranking, recommendations and A/B automation train on your catalogue, your shoppers and your real conversions, recorded per click. Personalisation uses a consent-gated profile, not a vendor's cross-client averages that may not fit your customers.

Question 9

How does Sellarix stop AI search from hallucinating or inventing products?

Accepted Answer

Sellarix only recommends products it actually pulled from your live catalogue, and explanations use real attributes, never invented ones. Guardrails check output for grounding and brand-safety before it reaches the shopper. No model is hallucination-free, so we prevent ungrounded answers by design.

Question 10

Is AI search and discovery GDPR compliant and privacy-respecting?

Accepted Answer

Yes, by default. Sellarix runs under GDPR and UK GDPR, provides EU AI Act Art. 50 disclosure, and offers UK and EU data residency with tenant isolation at the data layer. Long-term profiles are built only for consenting customers, and guest sessions store nothing against a person.

	Sellarix	Traditional / point tools
How a query is understood	Reads the full question: intent, synonyms, budget and use.	Matches keywords; compound and long-tail queries often miss or return nothing.
Architecture	Search, recommendations, visual, fit and bundling on one data spine, so signals compound.	Separate vendors that each learn only from their own slice of behaviour.
Explainability	Picks explained against real product details, kept grounded not invented.	Black-box scores, or chatbots that hallucinate when not grounded in live data.
Model and vendor lock-in	Model-agnostic: Claude, GPT, Gemini or local, swappable per task.	AI search often tied to one approach and gated to a top plan tier.
Agent readiness	Opt-in catalogue feed and agent-checkout API for emerging agent protocols.	On-site search only; little provision for external AI shopping agents.

AI product search and recommendations, on any platform

Overview

How it works

One shared data spine

Understands the whole question

A grounded, model-agnostic brain

Learn, test and convert

Capabilities

Use cases

Apparel: cut returns, lift baskets

Parts, electronics and DIY

Rescue searches you're losing

How Sellarix compares

Discover & Convert FAQ

Part of one platform

Bring this to your store