Plug Mira into
Platforms
Built for ecom
Get started
One widget on every product page, every help page and every checkout. The page decides which team answers, Mira handles order status and returns without a human, the support team overflows to the next team in 15 seconds, and the sale lands where it should. Across 50 stores and 98,000 conversations last year, WhatsApp chats converted at 40% against 5% on web chat.
Used by Veke, Interflora, Stockmann, Sareskoski and Masku.
EBEmma Berg<emma@demo.example>
to soporte@serviceform.com · cc Albert R., Ana C. +3
Hi team, following up on what Juli mentioned.
We can use the flows we already have with Automotor as a reference.
On Fri, 12 Jun 2026, Juli wrote:
GRUP balartFor both WhatsApp flows we've updated the capture flow for the sale.
— Ecommerce live chat benchmarks 2026
Twelve months of live chat across 50 online stores on Serviceform, aggregated and anonymised. Use it to set your own targets.
98 000
live chat conversations from 50 online stores, September 2025 to August 2026
40%
of WhatsApp conversations became a lead, against 5% of web chats. WhatsApp is 12% of chats and half of all leads
4.6×
higher lead rate when the chat starts on a product page: 21% against 5% on other pages
26%
of conversations were handled by the AI agent from start to finish, without a human
28%
of leads arrive outside 8 to 18 office hours, and one in five on a weekend
36%
of human-answered chats got a reply within a minute, median wait 90 seconds
What the data says
Source: Serviceform live chat statistics, 97,592 conversations from 50 ecommerce companies, 1 September 2025 to 31 August 2026. Aggregated and anonymised. A lead is a conversation where contact details were captured and emailed to the store. Wait time is measured from the shopper's first message to the first human reply on chats a human joined.
— Not a support desk chat
A support desk tool measures tickets resolved. A store measures orders. The chat has to know the product, look up the order, move the serious shopper to WhatsApp, and never leave a checkout question waiting for a second line that is at lunch.
Support desk chats
Every chat lands in one support queue, whatever page it came from.
Serviceform
The product page routes to the team that knows the product, returns go to returns, checkout goes to the team that can close it.
Support desk chats
A bot that asks for the order number and creates a ticket.
Serviceform
Mira looks up the order in Shopify or WooCommerce and answers delivery, changes and returns in the shopper's language, at any hour.
Support desk chats
The shopper waits in the queue while first line is busy.
Serviceform
Overflow in 15 seconds to the second line, then to the outsourced desk, in the same thread.
Support desk chats
WhatsApp as a separate app someone checks twice a day.
Serviceform
WhatsApp per brand in the same inbox, with product cards, carts and payment links. It converts at 40%.
Support desk chats
Conversations that disappear when the ticket closes.
Serviceform
Every shopper and every conversation saved to the customer data platform, with the products and the orders, for the next campaign.
40%
of WhatsApp chats become a lead, 5% on web chat
80%
fewer order enquiries reaching support (Masku)
26%
of chats handled by Mira without a human
15 s
until a chat overflows to the next support team
— How store routing works
Four things happen behind one widget: the page picks the team, the overflow chain picks who answers, WhatsApp follows the same map, and the shopper and the order end up in the customer data platform.
Rules read the product, the category and the cart on the page, so a chat on a sofa goes to furniture, a chat on a return goes to returns, a chat at checkout goes to the team that closes.
Each team has hours and a time zone. If first line does not pick up, the chat moves on in the same thread.
0 s
First line support
15 s
Second line
30 s
Outsourced desk
After hours Mira answers order status, delivery and returns from the order system and captures the lead for the morning.
Each store or brand has its own number, mapped to its team, with the same overflow chain behind it.
+358 40 …
Main store
+358 50 …
Outlet
A chat with contact details becomes a lead and a profile the moment it is captured, with the products and the order attached.
Resend to the warehouse or a partner store
A question on a sofa goes to the furniture team, a question on a return goes to returns, a checkout question goes to the team that can close it. The rule reads the product, the category and the cart, so the shopper never explains what they are looking at.
Where is my order, can I change the address, how do I return this. Mira looks up the order in Shopify or WooCommerce and answers in the shopper's language at any hour. Masku moved 80% of order enquiries off the support team this way.
fewer order enquiries reaching the support team
Masku connected the chat to its order system so shoppers get their delivery status instantly and the outsourced support team handles the rest.
First line is busy. After 15 seconds the chat moves to the second line, then to the outsourced desk. The shopper sees one thread. Nobody waits for a reply that never comes.
A WhatsApp number per store or brand, with product cards, carts and payment links in the thread. Identified shoppers convert at 40%, so the job of the web chat is to move the serious ones to WhatsApp.
AI summary and reply
Every conversation gets a three line summary and a suggested reply drafted from your quick replies, your own instructions and the product the shopper is looking at. Confidence is shown as a number. The agent adds it to the chat, edits it, or teaches the AI a better answer for next time.
Emma Berg
Web chat · live now
AI summary
Visitor came in through Quick contact and greeted Hiruni by name.
No question or request has been raised yet.
Company is still unknown, so the chatbot asked for it last.
Suggested AI reply Confidence 86%
Hi! Happy to help. Which company are you with?
WhatsApp for stores
A WhatsApp number per store or brand, each mapped to its team, with a first menu so the shopper picks order, product or return. Product cards, carts and payment links in the thread. Same inbox, same overflow chain, same customer profile as the website chat.
Nike Store
online
Do you have size 42 in this? 14:32 Built for stores and support teams
One store or a group of brands, an in-house team or an outsourced desk, the setup is the same. Teams per brand and topic, a live chat only login for the outsourced agents, one widget, one inbox, and every shopper in the customer data platform.
Teams per brand, country and language. Head office sees everything, each brand team sees its own queue.
A live chat only role for the outsourced support desk. They answer chats and see nothing else in the platform.
Order lookup, product data and carts from your store platform, so Mira and the agents answer with real stock and real delivery dates.
Every conversation, product and order saved on the shopper's profile, ready for the next campaign and the next chat.
Customer story · Masku
— Hosted in Europe
Your customer data stays where it belongs. We run on EU infrastructure, with no subprocessors outside the EU. Compliance is not a feature you turn on.
Every byte hosted in the European Union. Frankfurt, Helsinki, Dublin. Not "available in EU." Just EU.
Data processing agreement signed on day one. Right to erasure, portability and access wired into the platform.
No US subprocessors. No grey-zone analytics. Your customer data does not leave the EU.
— Store questions
Routing, order lookup, WhatsApp, overflow and outsourced support, answered plainly.
— Live chat by industry
Routing, overflow, leads and AI drafts set up the way your industry actually sells.
— Free analysis
Book a free 20-minute call with Daniel. We go through your store, your support team and your WhatsApp together, so you see exactly where sales are slipping away.
Real talk, no fuss. No slide decks, no pressure to buy.
