Every order question answered before it becomes a ticket

Where is my order, can I return it, does it come in blue. Synth answers with live order and stock data, and issues refunds and exchanges inside the rules you set.

The atrium of a bright shopping mall, with shopfronts on several floors

78%

78%

of contacts resolved without an agent

−41%

−41%

cost per contact in the first quarter

0.6s

0.6s

median reply on chat

4 wks

4 wks

from contract to live traffic

The conversations that fill a retail queue

Seven intents make up almost every retail contact. Each one maps to a flow that ships with the retail pack, and you can change any of them.

Intent

Share of contacts

Resolved

Flow

Where is my order

“It said Tuesday and it is Thursday”

34%

92%

resolved

Order status

Order status

Returns and exchanges

“Can I swap this for a medium?”

21%

81%

resolved

Returns

Returns

Product questions

“Is the jacket waterproof or just resistant?”

15%

88%

resolved

Catalogue answers

Catalogue answers

Delivery changes

“Can you leave it with a neighbour?”

11%

74%

resolved

Delivery update

Delivery update

Payments and promo codes

“My discount code did not apply”

9%

69%

resolved

Payment check

Payment check

Account and loyalty

“Where did my points go?”

6%

77%

resolved

Account lookup

Account lookup

Everything else

“Complaints, press, the unexpected”

4%

To a person

Hand-off

Hand-off

The model writes the reply. The flow moves the money.

Returns run as fixed steps you can read. The language model understands the customer and writes in your voice, but whether a refund is allowed is decided by a rule, every time.

Return windows, final-sale items and refund limits set per market

Exchanges checked against live stock before they are offered

Labels created with your carrier, refund released on first scan

Anything outside policy handed to a person with the case attached

A woman packing an order into a cardboard box beside a rail of colourful shirts

The jacket is a bit small. Can I swap it for a medium?

Exchange placed

Inside the window, medium in stock

Black Friday without the hold queue

Contacts tripled across peak week at one fashion retailer. The agent took the extra volume and the support team stayed the size it was in October.

19,600contacts on Cyber Monday, 82% resolved by the agent

  • All contacts
  • Handed to a person
Contacts per day across peak week
DayContactsHanded to a person
Nov 216,2001,400
Nov 225,8001,300
Nov 237,1001,600
Nov 248,4001,800
Nov 259,9002,100
Nov 2611,6002,300
Nov 27, Black Friday18,9003,400
Nov 2816,2003,000
Nov 2914,1002,700
Nov 30, Cyber Monday19,6003,600
Dec 115,3002,900
Dec 212,2002,400
Dec 310,1002,100
Dec 48,7001,900

Connected to the systems that hold the answers

The retail pack comes with connectors for the usual stack. Each one lists exactly what the agent can read and what it is allowed to change.

System

Reads

Writes

Storefront

Reads

Reads

Catalogue, prices, stock by size

Writes

Writes

Nothing

Order management

Reads

Reads

Orders, fulfilment, tracking

Writes

Writes

Delivery notes, exchanges

Payments

Reads

Reads

Charges and refund status

Writes

Writes

Refunds inside policy

Returns portal

Reads

Reads

Open returns and labels

Writes

Writes

New returns, labels

Loyalty

Reads

Reads

Points, tier, history

Writes

Writes

Goodwill points, capped

Helpdesk

Reads

Reads

Past tickets

Writes

Writes

Hand-offs with a summary

Portrait of Priya Raman

parcelry

“We used to review fifty tickets a week and hope they were typical. Now every conversation has a score, and the first thing we fixed was a flow we thought was fine.”

Priya Raman

VP Customer Operations, Parcelry

−38%

handle time in the first quarter

Questions retail teams ask

Which commerce platforms does Synth connect to?

Can the agent issue refunds on its own?

How does it handle peak season?

Does it answer product questions from our catalogue?

Put Synth on your retail queue

Send us a month of transcripts and we will show you which conversations an agent could have finished, and how.

Put Synth on your retail queue

Send us a month of transcripts and we will show you which conversations an agent could have finished, and how.

Product notes, once a month.

What shipped, what we measured, and what we got wrong. No tracking pixels.

Sign-up is off in this preview. Connect a form endpoint in the site config to turn it on.

© 2026 Synth. All rights reserved.

Create a free website with Framer, the website builder loved by startups, designers and agencies.