Blog
Articles on how the support agent works and where its limits are.
Customer service statistics 2026: 30 figures every store owner should know
63 percent switch after one bad experience, 88 percent expect faster replies than a year ago. 30 documented customer service figures and what they mean for your store.
Black Friday customer service: the 3-week checklist
On Black Friday the orders arrive in waves, and the support requests follow two days later. What you get done in the three weeks before decides how your December goes.
AI chatbot vs. AI agent in support: the difference that costs money
Both answer customer questions, both have “AI” in the name. Only one of them can tell you where the parcel is. The difference costs you hours every month.
Reduce your return rate: 9 measures with a measurable effect
Most returns have three causes, and none of them is the customer. Nine measures that address the causes, with the number you use to measure success.
The 12 most common customer requests in e-commerce – and how to standardise each
Almost every support email falls into one of twelve categories. If you know them, you have the answer for ten of them before the email arrives.
Shopify customer service apps compared: Inbox, Gorgias, Zendesk, Humane
Four tools, four approaches. We compare Shopify Inbox, Gorgias, Zendesk and Humane the way you would when choosing: by the problem you have right now.
Customer service software pricing: why we bill per ticket, not per seat
Per seat, per feature, per resolved conversation, or per ticket. The billing model decides how a support tool feels day to day. Our position, with the math.
Return rates in Germany 2026: figures by industry
One in four parcels in German e-commerce comes back, and far more in fashion. The documented return rates by industry, with costs and context.
“Where is my order?” – reducing WISMO requests before they happen
The most common support request in an online shop is not a question for you but for the carrier. Here is how to bring it down to a fraction.
Returns process for online shops: step by step from request to refund
A return is not one event but a chain of six steps. Define each one properly once and every return after that costs you less time and fewer emails.
AI in customer service: what actually works in 2026
Somewhere between “the chatbot that annoys everyone” and “support runs itself” is a zone that genuinely pays off for small shops. This guide shows where it starts and where it ends.
Customer service for online shops: the complete 2026 guide
Everything support means in a small Shopify shop, in one text: from the first shipping email to deciding what you automate and what you don't.
Train the AI agent once instead of correcting every reply
Tone, goodwill limits, trigger words. Set once, not explained per email.
Support metrics that tell you where the time goes
First contact resolution, autosend share, post-processing rate. And which one to read first.
A returns portal instead of five emails for one parcel
The customer files the return himself, with a reason and a photo.
One ticket instead of three tabs across Shopify, tracking and inbox
The order, the payment and the shipment sit next to the message, not in two other tabs.
Automatic replies for everything that arrives after hours
Four conditions have to be true at the same time, or nothing goes out.
When humans stay in the loop
Automation for speed, people for judgement. Where the line runs, and who draws it.
GDPR in day-to-day support, not just in the imprint
Encryption, German servers, a processing agreement. And what of that is visible in the product.
Email support that grows with you
A clean intake turns an inbox into cases you can actually work.
What counts as a ticket, and what does not
A ticket is a customer request, not every single email. Where the line sits.
Shopify context in support, without anyone hunting for order numbers
The agent reads the order, the payment and the shipment before it writes.
Why we do not train on your customer conversations
What your agent learns, it learns for you. What that means technically, and what it rules out.
How we review AI replies before they reach your customers
Every draft has to earn the send. What gets checked, and when a human takes over.