Online brands · Shopify and similar · 500+ orders a month · sale and Black Friday peaks
“Where's my order?” answers itself. Your team keeps what actually needs a human.
The agent looks up the real order, says where it is, handles exchanges and returns, recommends from your catalog, and sells in the chat. In seven languages if you need it. And it hands your team whatever falls outside your rules.
Ask Sol about an order or ask for a recommendation.
The bookshelf arrived with a cracked shelf 1:12 PM ✓✓ 
Aalto table · oak 32 × 20 in · in stock $219 Add to cart 
Lumi table · ash 36 × 18 in · in stock $269 Add to cart 
Kivi table · birch 28 × 28 in · 3 left $289 Add to cart Dashboard · today
What slips away today
What gets lost every week without anyone noticing
01
Half your support is “where's my order?”
And a good chunk of the rest is exchanges and returns. Questions your Shopify already answers that today cost a ticket, a person, and a wait.
02
Peaks break the service
Sales, Black Friday, the holidays: tickets triple and response time explodes right when you sell the most.
03
Nobody's selling in the chat
Whoever asks about a size, a dimension, or a color at 9 PM is about to buy. No answer, they leave. And nobody counts how many.
The journey
What happens when a customer texts, from order to repeat purchase.
The agent acts in your actual store: looks up, changes, generates labels, and takes payment. And it knows how far it's allowed to go.
-
01 Asks
Order, shipping, return
Looks at the real order in your store and the status with the carrier. Answers with data, not “24-48 hours”. And if the customer comes from an ad or a button on your website, it starts with that context.
-
02 Resolves
Answers, not excuses
Applies your exchange and return policy, opens the case with the photos, and leaves it ready for your team to confirm.
-
03 Sells
Recommends from your real catalog
Stock, sizes, dimensions, compatibility. Cards with photo and price, and payment in the chat.
-
04 Updates
Back in stock
“Your size is back in stock”, with the customer's opt-in and no spam.
-
05 Escalates
A person for anything off-script
Refunds outside policy, repeat complaints, anger. With the full history and priority.
-
06 Measures
Tickets avoided and chat sales
What resolves on its own, what goes to a person, how much sells in the chat. Every month, with a name on it.
The life of a conversation
The states every conversation goes through, when a person steps in and how it closes.
What the agent does here
What it gets done, playing by your rules
Orders
Sees the real order and the carrier
Status, tracking, ETA, issues. Connected to Shopify and your carrier.
Returns
Exchanges and returns without forms
The agent applies your policy, collects the photo, and opens the case; the label and the refund are confirmed by your team.
Catalog
Recommends what you actually have in stock
Understands dimensions, sizes, materials, and compatibility. Never recommends what's sold out.
Sales
Takes payment in the chat
Cart and checkout in the same thread, plus a note when what they wanted is back in stock.
Handoff
Anger or exception → a person
Detects frustration and requests for a human from the first message and transfers with priority.
Metrics
Tickets avoided and chat sales
Resolution, response time, assisted sales, and escalation reasons. By language and by channel.
A day with the system
An ordinary day at a business like yours, told by the channel itself: what happens while the team works, and while it sleeps.
8:00 AM
The team opens the inbox and finds twelve overnight conversations already resolved: orders located, sizing questions answered, and two cases opened with their photos, ready to confirm.
11:40 AM
A customer writes in angry for the third time. The agent doesn't run her in circles: it hands her to Marta with priority and the full history, and Marta fixes it in four minutes.
1:12 PM
A photo of a cracked shelf comes in. The agent sees it, applies your policy, and leaves the case open with the photo inside. No form, no ticket.
7:05 PM
Someone's looking for a table for a small living room. The agent shows three in stock with photo and price and closes the sale with shipping included.
11:30 PM
Black Friday peak: 300 conversations in an hour, in four languages. All answered. The team sleeps.
-
8:00 AM
The team opens the inbox and finds twelve overnight conversations already resolved: orders located, sizing questions answered, and two cases opened with their photos, ready to confirm.
-
11:40 AM
A customer writes in angry for the third time. The agent doesn't run her in circles: it hands her to Marta with priority and the full history, and Marta fixes it in four minutes.
-
1:12 PM
A photo of a cracked shelf comes in. The agent sees it, applies your policy, and leaves the case open with the photo inside. No form, no ticket.
-
7:05 PM
Someone's looking for a table for a small living room. The agent shows three in stock with photo and price and closes the sale with shipping included.
-
11:30 PM
Black Friday peak: 300 conversations in an hour, in four languages. All answered. The team sleeps.
What gets measured
The numbers that go into the monthly report
−58%
“where's my order” tickets that reach the team
24/7
support in seven languages, Black Friday included
+19%
assisted sales in the chat on the same traffic
Illustrative figures of what the system moves: your pilot's baseline and target get agreed on your data, not these.
Impact calculator
How much is slipping away today?
~400
Conversations without a fast reply each month
With a first reply in seconds, those ~400 conversations stop waiting. How many end in a booking or a sale depends on your business: that is what the pilot measures.
Compliance in your industry
What legal will ask, handled for your case
Returns and warranties per the law and your policy
The agent applies your written policy and the return rules that apply to you, and promises nothing outside it. Anything beyond goes to a person.
Order data, nothing more
It only touches what's needed to help: order, shipping, purchase history. Dedicated infrastructure, a signed data-processing agreement, and no training models on your customers. CCPA handling is mapped in the diagnostic.
It says it's an AI and steps aside for a person
From the first message. Requests for a human are detected instantly and time-to-human is measured on the dashboard, not promised.
What it connects to
It connects to your store, your carrier, and your helpdesk.
Shopify is in production with a real brand; WooCommerce and carriers with an API get connected in the diagnostic. It coexists with your helpdesk or replaces it: your call.
We already have Gorgias with macros and an FAQ bot.
A macro answers; this works: it looks up the order in your store and the status with the carrier, applies your return policy, opens the case with the photos, and takes payment in the chat, with rules in code an FAQ bot doesn't have. Anything touching money or addresses is confirmed by your team. It coexists with your helpdesk or replaces it, we decide that with you.
Does it speak our customers' languages?
Yes. The brand we have in production serves seven languages with the same agent, detecting the customer's language from their first message.
What if it gets a refund wrong?
It can't: refunds and exceptions to your policy sit in the “only with confirmation” or “always a person” lists. Those rules live in code, outside the model.
How does it hold up on Black Friday?
Hundreds of conversations at once with no queues and no temp staff. Your team only steps into what's off-script. In the diagnostic we size it with your peak history.
Other industries
Your industry? See it on your own cases.
Bring three real conversations from your WhatsApp and we'll replay them in the demo.