Changelog

This is our first general-availability release. Cartfilo turns a shopper’s plain-language request — “a snack basket under 500,” “running shoes under 2000 that aren’t too flashy” — into a ready-to-buy cart, right inside your storefront chat. Since our earliest builds we’ve shipped budget-aware multi-item baskets, conversion and revenue numbers calculated from real orders (not estimates), clear in-dashboard notice whenever a plan limit changes answer quality, returning- customer recognition, and a faster human handoff for the moments a shopper needs a real person. Thank you for being here from day one — this is just the start.

Your team now gets an email the moment a shopper needs a human, instead of finding out only if someone happens to have the inbox open. Saved reply templates speed up your most common answers, and a handoff now stays attached to a shopper even if their session refreshes mid-conversation — no more losing track of who’s already talking to a real person. Live updates in the inbox are also snappier and stay scoped to the conversation you’re actually viewing.

When a shopper shares their email — during checkout or while talking to a member of your team — Cartfilo now matches them to their real customer record, so your staff can see who they’re talking to instead of an anonymous visitor. Obviously fake, disposable email addresses are quietly flagged at the door. And if a shopper ever asks you to export or delete their data, that request is now fully automated end to end.

Free-plan stores that reach their monthly message cap now see a clear, dismissible banner on their dashboard explaining exactly why answer quality changed, with real usage numbers and a direct link to upgrade — instead of a silent quality drop with no explanation.

Your dashboard’s Conversion Rate and Revenue Influenced figures are now calculated from real Shopify order data tied back to actual chat sessions — not estimates. Revenue is split into a directly-attributed figure and a separate estimated figure, so the headline number never overstates what the assistant actually influenced.

Requests like “snack basket under 500” now respect the exact number a shopper typed when assembling a multi-item basket, instead of only applying budget constraints to single-item searches.