Every label sizes differently
A large in one brand is a medium in another, and "runs small" in a review means little without numbers. A shopper who can't match a hoodie they own to your chart guesses, orders two sizes or leaves.
A size chart can't ask what a shopper wears in another brand, and a call can. When a checkout stalls, a Loop Voice agent compares the shopper's usual size with your garment measurements and fit notes, quotes a drop's ship date as published and texts a link back to the cart, without inventing scarcity.
The short answer
An AI voice agent for an apparel brand phones shoppers who abandoned checkout, settles the size question from your chart and fit notes, quotes ship dates and exchange terms as published, and sends the cart back by text. Low-stock lines stay out of its script, because a count that was true when you loaded it may not be true when it calls. Loop Voice hasn't published apparel numbers in its September 2026 benchmark, so your first weeks of calls set the baseline.
A large in one brand is a medium in another, and "runs small" in a review means little without numbers. A shopper who can't match a hoodie they own to your chart guesses, orders two sizes or leaves.
Preorders and timed drops sell a ship date as much as a garment. Shoppers want to know when the order leaves the warehouse and what happens if that date slips.
A streetwear cart is often one hoodie, so the whole shipping charge lands on a single piece. A call can say how far the cart is from your free-shipping threshold without pushing a second item.
Free exchanges, final-sale drops and who pays return shipping all change the decision, and those terms sit on a returns page, not next to the size selector.
When a checkout stalls on a size, the call answers it and a text carries the cart link back; the flow ends the moment an order lands. If single low-priced tees fill your abandoned checkouts, a cart-value branch keeps calls for larger carts and sends the rest a text.
See the use caseShoppers who signed up for a sold-out size have already chosen the item. When it lands, run a campaign (in beta) to a Klaviyo or Shopify segment of the ones who agreed to calls, or upload that size's waitlist as a CSV to start a flow. Klaviyo logs a Subscribed to Back in Stock event on each profile that signs up for an alert.
See the use caseOn a drop day, order status, size and exchange questions arrive at once. The inbound agent takes them around the clock and opens a ticket when an exchange needs a person.
See the use caseA short call a week after delivery asks where the size runs tight or loose. Sample a share of orders with a split step and feed the answers into your size chart.
See the use caseAI agent
Hi Jordan, it's Alex, an AI assistant calling from [Store]. If you'd rather not get these calls, just say stop. You left the heavyweight hoodie in your cart. Was it the size?
Shopper
Yeah. I'm a large in most brands, but the reviews say yours run big.
AI agent
It's cut boxy with a dropped shoulder. Lay a hoodie you like flat and measure armpit to armpit, and I'll match it to the size chart.
Shopper
About 23 inches.
AI agent
The medium is 23.5 inches across and the large is 25, so the medium is closer. Your cart has the large, so swap it before you check out.
Shopper
And if it's still wrong?
AI agent
Our returns page says size exchanges are free within 30 days. Want me to text you a link back to your cart?
Shopper
Yeah, send it.
What shoppers say
“I'm a large everywhere else. What am I in yours?”
How the agent answers
It asks for a measurement from a garment the shopper owns, compares it with your size chart and names the closer size. If the chart has no measurement for that style, it says so instead of guessing.
What shoppers say
“Will it sell out if I wait till Friday?”
How the agent answers
It doesn't guess, and it doesn't use a sell-out as pressure. With the create urgency objective off, it points the shopper to the size selector on your site and leaves the timing to them.
What shoppers say
“When does the preorder actually ship?”
How the agent answers
It quotes the ship window on the product page and never an earlier date. Asked what happens if the date slips, it explains your delay policy, including the prompt-refund option the FTC's Mail Order Rule gives buyers when an order can't ship on time.
What shoppers say
“Can I return it if it doesn't fit?”
How the agent answers
It reads your exchange and return terms, including final-sale items and who pays return shipping, before the shopper pays. Refund disputes and exceptions are handed to your team in a ticket.
What shoppers say
“Can you hold a medium for me?”
How the agent answers
It answers from your policy. If you don't hold stock for anyone, it says so and offers to text the cart link, so the shopper can check out while the size is there.
Start in Settings, Integrations by connecting Shopify, WooCommerce or Checkout Champ. An abandoned checkout starts the flow, an order stops it, and the agent sees the style and size in the cart.
Put a size chart for each style in the knowledge base: garment measurements in inches, fit notes such as boxy or cropped, and your exchange and final-sale terms. Product documents refresh nightly, so re-crawl the day a drop page goes live.
Turn on the abandoned checkout flow and decide whether a drop carries a code at all; if you allow none, the agent offers none.
Leave the create urgency objective off, and write a custom guardrail: never state a stock level, and never a ship date the product page doesn't show.
Restock lists need their own consent check: a back-in-stock signup covers AI calls only if its wording says so, and the calling rules guide explains what written consent needs. Calls go out within the hours you set for the store, in the shopper's own time zone.
In the simulator, write a one-off shopper who's between a medium and a large and sure the hoodie runs big. After launch, check the QA notes for styles that draw size questions.
It can say which size is closer on your chart, and the choice stays with the shopper. Garment measurements for each style work better than a body-measurement table, because a shopper can lay a hoodie they own flat and compare. For shoppers between sizes, load fit notes such as "size down for a closer fit", and the agent reads them out.
Usually not. Each answered conversation costs roughly $0.95 to $1.49 when billed monthly, per the cost calculator, while unanswered calls use no credits. A cart-value branch, with a minimum Loop Voice suggests from your results, sends small carts a text instead.
Yes, as an outbound campaign (in beta) to a Shopify or Klaviyo segment, with your own daily window and calls-per-minute pace. For most drops a text is enough; save calls for customers who asked about a specific item and agreed to AI calls.
Update the date in the knowledge base the day it changes, so the agent quotes the new one. Under the Mail Order Rule, you have to give the buyer the choice of waiting or canceling for a prompt refund, and the agent can explain both when asked.
It explains your exchange terms, looks up the order and files a ticket for the exchange, copied to Gorgias, Zendesk or your support inbox if connected. Call back within 7 days and the agent knows the story.
Count it. For the first few weeks, search the call records for size and fit; each call's QA notes list the shopper's objections. Loop Voice's September 2026 benchmark mixes every kind of store, and across them a question about the product (size, fit or compatibility) was the reason given in 5.5% of 824 abandoned-checkout conversations. With no apparel cut, your own count, next to answer rate and attributed orders per flow, is the one to plan around.
Install Loop Voice and start generating revenue in under 10 minutes.