Your search bar can’t close the sale. Chat can.
Listed on the Shopify App Store as Lagrange AI Shopping Assistant.
Recommendation · in stockSalomon Speedcross 6 · $140 
Better fit · wide toe boxAltra Lone Peak 8 · $150 Order #48213 · updatedSwapped to US 10 · arrives Fri - What is the AI Shopping Assistant?
- The AI Shopping Assistant is a Webscale AI product that sells and serves on your storefront in one conversation. It answers from the Commerce Context Engine, which combines your catalog, back-office systems and expert knowledge. It also writes orders, exchanges and returns back to your systems.
Answers that close the sale
A shopper explains their setup in plain language. The assistant answers from the spec sheet, an expert interview and live inventory, recommends the right item and completes the setup in the cart.
Setup, platform and space, in plain language.
Spec sheets, expert knowledge and live inventory.
Items that work together, not just the one searched for.
No search, no dead ends.
Everything a shopper needs, in one thread
A single question can span discovery, product details, order status and returns. The assistant routes each part to the right specialist and answers as one, so shoppers never repeat themselves.
Product discovery
Reads intent in plain language and takes shoppers straight to the right item, no filters or keyword roulette.
Product Q&A
Answers sizing, materials, compatibility and shipping, sourced from your attributes and docs.
Recommendations
Compares options and suggests complements from the conversation, not last-click retargeting.
Order status
Looks up orders and tracks shipments in the same chat, no ticket and no login maze.
Returns & exchanges
Starts returns, checks eligibility against your policy and offers exchanges before the sale is lost.
Account-aware
Recognizes signed-in shoppers and dealers, then reflects contract pricing, history and entitlements.
When a conversation needs a human, it hands off with full cart and chat context intact, no starting over.
AI assistant
Advanced Settings
ResetSaveGeneral
SaveStyling
PreviewSaveIllustrative. Specialist agents in the AI Shopping Assistant settings.
Webeyez sees the friction. The assistant steps in.
The AI Shopping Assistant integrates with Webeyez, the session-intelligence platform. Webeyez reads every live session and fires a signal the moment a search fails, a page stalls, or a shopper is about to leave. The assistant opens in that same session with a real answer instead of letting the moment pass.
search_no_resultspage_latency_highunproductive_browseexit_intentidle_on_fitmentcustom_triggerSignal layerWebeyez reads every live session in the background and names the break the moment it happens, then fires that signal onward.
The AI Shopping Assistant opens on the storefront in that same session, asks what the shopper actually needs, and hands them a real product.
Every storefront has its own friction points. During onboarding, we build the exact triggers yours needs, not just the ones above.
Live on your storefront in 3 steps
Connect your store with no code and no replatforming, let Forward Deployed Engineers ground the assistant in every source, and let shoppers talk.
Connect your store
A no-code install on Adobe Commerce, Magento, Shopify, Shopware, Salesforce Commerce Cloud or BigCommerce. The Commerce Context Engine starts pulling in your catalog, pricing and policies.
Ground it in every source
Forward Deployed Engineers bring in spec sheets, back-office data and expert knowledge, then normalize and validate it, so every answer comes from the system that owns the fact.
Let shoppers converse
Discovery, Q&A, order status and returns run in one thread. You watch it work and tune guardrails from a single console.
AI assistant
Install Tags
SaveAdd this code to your website’s main template, before the </body> tag. It will automatically be added to every page on your website.
<script src="https://cdn.webscale.ai/js/companion-widget.js" data-agent-code="YOUR-AGENT-CODE" data-merchant-id="YOUR-MERCHANT-ID" defer ></script>Copy Step 2
Confirm onsite tracking is enabled
Illustrative. The install tag in the AI Shopping Assistant settings.
Plus connectors for Zendesk, Gorgias, Freshdesk, Klaviyo, your OMS and knowledge base, so the assistant answers with full context.
Sounds like your best associate
Before launch, our team captures how your brand should sound from brand guides and interviews with your experts. You set what the assistant says and what it never says, and Forward Deployed Engineers keep tuning it on real conversations.
From brand guides and expert interviews.
What it says, and what it never says.
Engineers review conversations and refine.
One voice on your storefront, in AI chat ads and off-site.
Accuracy built in, not bolted on
In compliance-bound categories a wrong answer is a liability. You own the orchestration layer, and we build the guardrails and knowledge bases with you.
Grounded & cited
Every answer is sourced from your live catalog, your product reviews and your approved knowledge base, with citations behind the scenes.
Guardrails for regulated verticals
Jurisdiction logic and policy rules keep answers compliant where a mistake actually matters.
Hallucination controls
If the assistant does not know, it says so and offers a path to a human, protecting your brand.
Auditable chat logs
Every conversation is logged and reviewable, so compliance and support can read back exactly what the assistant said and what it sourced.
Logs
Export| Time | Conversation | Sources | Feedback |
|---|---|---|---|
| 14:52 | Does the 2023 slip-on fit my bike, and is it in stock? | 3 cited | |
| 14:49 | Where is my order? It said Tuesday. | 2 cited | |
| 14:47 | Can I exchange these for a half size up? | 4 cited | |
| 14:44 | Which tent is better for high wind, the Ridge 2 or the Peak 3? | 3 cited |
Illustrative data. Conversation logs in the Webscale AI workspace.
Turn it on for as little as you want
You choose who sees the assistant before anyone outside your team does, and you can widen or withdraw that scope at any point without a release.
Internal users only
Your team sees the assistant on the live storefront. Nobody else does.
QA only
Run it against real catalog and order data in a controlled environment before it meets a shopper.
Specific cohorts
One category, one region, one account tier, or a named list of B2B accounts.
A percentage of traffic
Start at a few percent of sessions and widen as the transcripts hold up.
Every scope is reversible from the same console, and every conversation in it is logged for review. If something is wrong, you pull it back before it reaches the next shopper.
One price to put the assistant on your store
Start on the AI Shopping Assistant, or roll it into the full platform with Audience Segmentation when you are ready. No replatforming, no lock-in.
- $0.98 per conversation
- Conversational discovery, Q&A and support
- Grounded in your live catalog and knowledge base
- Order, returns and product agents included
- Account-aware pricing and history
- Standard connectors and live-agent handoff
The short answers
From the Commerce Context Engine: your catalog and site content, reviews, orders and returns, inventory and pricing, spec sheets, support history and interviews with your experts. Forward Deployed Engineers normalize it before launch.
Yes. It finds the order, checks eligibility against your policy and creates the exchange or return in the conversation. The change writes back to your OMS and returns portal.
The connection installs in under a day with no code. Launch timing depends on the state of your data. We launch a focused first version, then phase in more. A small merchant with a messy catalog took about eight weeks of data work.
Yes. You can scope the assistant to internal users, to QA, to specific shopper cohorts, or to a percentage of live traffic, and change that scope at any time. Every rollout step is reversible, and every conversation is logged so you can read back what was said before you widen it.
No. Answers are grounded in your live catalog and approved knowledge base, with hallucination controls on. If the assistant does not know, it says so and offers a path to a human.
Adobe Commerce, Magento, Shopify, Shopware, Salesforce Commerce Cloud and BigCommerce, with connectors for tools like Klaviyo, Zendesk and your OMS so the assistant answers with full context.
Yes. When a conversation needs a person, the assistant hands off with the full cart and chat history intact, so the shopper never has to start over.
Yes. You own the orchestration layer, and we build guardrails and knowledge bases with you so answers stay accurate and compliant, with jurisdiction logic where it matters.
One conversation that sells and serves
Connect the Commerce Context Engine and launch the AI Shopping Assistant on the store you already run. The same connection reaches every other channel when you are ready.