In many shops, customer service is already digital but spread across too many windows. This guide shows you which channels matter, in which order to bring them together and where automation reaches its limits. The figures come from our own benchmark studies.
What is digital customer service?
Digital customer service means you receive customer enquiries via email, WhatsApp, Instagram, live chat and marketplace inboxes and answer them in one shared system. For an online shop, it works well when all channels flow into one inbox and your team sees the order data directly in the ticket. That way you answer standard questions without searching the shop backend and have time for the cases that need a personal touch.
It is also called digital customer care or online customer service. The difference from a traditional setup shows in your team's daily work:
| Traditional setup | Digital customer service | |
|---|---|---|
| Channels | Phone and a shared Outlook inbox | Email, WhatsApp, Instagram, live chat and marketplaces in one inbox |
| Order data | In the shop backend, copied and pasted into the reply | Directly in the ticket from Shopify, Shopware, JTL or Xentral |
| Same customer, two channels | Two separate cases, often two answers | One customer profile with the full history |
| Availability | Office hours | Standard questions in the evening too, once you automate replies |
| Reporting | Gut feeling | Enquiries per channel, response time, resolution time |
Digital means: all channels in one system. If your team answers Instagram DMs in the app next to the inbox, you are running two separate workstreams. Only the shared inbox with order data turns this into a service you can manage.
Why is digital customer service worth it for your shop?
Digital customer service pays off as soon as enquiries come in through more than one channel. It saves your team duplicate replies and the search for order data.
The market sets the scene: your customers buy online, and a large share of them via marketplaces. According to the HDE Online Monitor 2026 from the German Retail Federation, 56.7% of German online revenue in 2025 went through marketplaces. For 2026, the federation expects online retail to grow by 4.3% in nominal terms, compared with 1.6% for bricks-and-mortar retail.
For your service, this means: enquiries arrive through more routes than your team can keep track of. Picture the day after a discount campaign at a D2C shop that usually gets 150 tickets a day:
- 220 emails about delivery times and discount codes are waiting in the inbox.
- A colleague answers 40 Instagram DMs on her phone on the side.
- Nobody looks at the Amazon messages in Seller Central until the afternoon.
- A customer asks about the same parcel by email and by WhatsApp.
Each of these channels is manageable on its own. Together they create duplicate work. Two colleagues answer the same parcel question while a third searches for the tracking number in the Shopify backend. The customer gets two different answers and writes a third time.
Digital customer service brings these routes together. We explain the role AI plays in detail in our guide to AI in customer service. This guide covers the foundation underneath: one inbox, order data next to the message and the enquiries that cost the most time.
Which channels belong in digital customer service?
Email carries most service enquiries. Instagram and WhatsApp together account for just over a quarter. Ticket volume in shops with AI-supported service splits like this:
| Channel | Share of ticket volume | Typical enquiries | What to watch |
|---|---|---|---|
| 63.0% | Order status, returns, invoices | Contact forms arrive as email in many setups | |
| 15.4% | Product questions, availability, replies to Stories | DMs belong in the inbox rather than on a personal phone | |
| 11.2% | Parcel status, returns, questions about campaigns | Service replies run within the 24-hour window after the customer's last message | |
| 5.1% | Messages about ads and posts | Keep comments and direct messages in view separately | |
| Live chat widget | 2.2% | Questions shortly before purchase | Worth it if you reply quickly |
| TikTok | 1.7% | Questions about products from videos | Relevant once you sell via TikTok Shop |
| Marketplaces | No separate value in the study | Order status, returns | Each marketplace has its own rules |
With marketplaces, the details matter. In armincx by Chatarmin, you connect Amazon with order data and AI actions. OTTO and Kaufland run into the same inbox as an email channel. There, your team answers the messages without AI actions.
In practice: if you only answer the email inbox today, you cover just under two thirds of enquiries. The remaining third comes in via Instagram, WhatsApp, Facebook, chat and TikTok. Without a shared inbox, that is where duplicate tickets quickly appear.
The channel figures come from the AI Customer Service Benchmark: 131 shops, 2.9 million resolved tickets, June 2025 to June 2026. The sample shows shops that use armincx, so it is one slice of e-commerce. If you sell a lot via Amazon or OTTO, your channel mix looks different.
How to digitalise your customer service in 5 steps
If you want to digitalise your customer service, start with a count. The order below saves you from the classic mistake: five channels open and no overview.
1. Count your enquiries for two weeks
For ten working days, tag every enquiry by channel and topic. A simple table is enough, for example:
| Date | Channel | Topic | Minutes | Resolved without follow-up |
|---|---|---|---|---|
| 3 Nov | Where is my order? | 4 | yes | |
| 3 Nov | Size advice | 7 | no | |
| 3 Nov | Return | 6 | yes |
Then you can see in black and white which five topics fill your inbox and which channel eats up the most time. "Where is my order?" is often at the top. This list shapes every next step.
2. Bring all channels into one inbox
Every customer gets one profile with her full history. Once email, Instagram, WhatsApp and live chat flow into one inbox, you see straight away that last night's customer is following up via WhatsApp today. Your team replies once instead of twice.
Decide who handles which channels. From day one, answer DMs only in the inbox so that no conversation gets stuck on someone's phone. Also check that the contact form and marketplace inboxes flow into it too, as these routes are easy to miss. Our omnichannel inbox shows what this looks like in practice.
3. Connect your shop and inventory system
Digital service only gets fast once order data sits in the ticket. Connect your shop system and inventory management, for example Shopify or Shopware with JTL, Xentral or Billbee. Your team then sees order status, tracking number and invoice right next to the message.
The same enquiry before and after the connection: A customer writes via WhatsApp: "Where is my parcel?"
- Before: A colleague looks up the order in Shopify, copies the tracking number and replies. A second has already answered it by email.
- After: The inbox recognises the customer, shows the order with the tracking link in the ticket and merges both messages into one conversation. One reply is enough.
Check where the tracking number actually lives. If it is only in JTL, JTL needs to be connected. If your shipping runs via Sendcloud or a fulfilment provider, the parcel status comes from there.
4. Build answers and a knowledge base
Write one checked template answer for each of your ten most common topics. Add return periods, delivery times, size guidance and payment methods to a knowledge base. An example: "Your order leaves our warehouse within 24 hours. You will receive the tracking link by email as soon as DHL scans the parcel."
The AI uses this knowledge base later. Whatever is missing or outdated here, it will answer incorrectly. So have someone review the answers on a fixed date each month.
5. Set your metrics before you start
Measure a baseline before the switch so you have something to compare later. Three values are enough to start:
- Time to first response
- Resolution time until the ticket is closed
- Share of tickets resolved without manual handling
An example of a baseline: first response by email after 8 hours, resolution after 2 days, no ticket without manual handling. Record the values per channel. After the switch, you will see whether WhatsApp tickets close faster than email or a channel lags behind.
Formulas and comparison values are on our page with benchmarks for your service metrics.
How much does digital customer service resolve automatically?
Counted across all tickets, one in three runs automatically, through AI or workflows. The range between shops is wide:
| Metric | Value |
|---|---|
| Automated tickets across all 2.62 million tickets (December 2025 to June 2026) | 33.2% |
| Typical shop (median) | 16.3% |
| Average per shop | 25.2% |
| Top 25% of shops (average) | 61.2% |
| Tickets resolved entirely by AI alone | 4.7% |
| Monthly values | 24.8% (April 2026) to 38.5% (February 2026) |
According to the study, the rate depends heavily on setup, industry and ticket mix. Our view: within the setup, steps 3 and 4 matter most, meaning connecting your shop and inventory system and a well-maintained knowledge base.
For context: AI alone fully resolves only around 5% of all tickets. Most of the automation comes from workflows. The rate also fluctuates: in April 2026 it was 24.8%, with 1.9 times as many resolved tickets as in February. So plan with a range and calculate conservatively for your shop.
Translate the range to your shop. At 300 tickets a day, 16.3% means around 49 tickets handled without manual work. At 61.2%, it is around 184 tickets. The difference of 135 tickets a day is work your team can get back.
This calculation only works if the automated answers are correct. In the first weeks, spot-check every tenth automated answer before you add further topics. If the spot checks run for two weeks without errors, add the next topic from your initial count.
All values come from the Automation Benchmark with data from 140 shops.
Which enquiries can you have answered automatically?
Automated answers pay off for topics with a clear answer from your data. In daily shop life, that means six topics: order status, returns, cancellation before dispatch, invoices, address changes and product advice from your knowledge base.
All six need a connection so that the answer is right. A chatbot without access to the order can only point to a form. With access, armincx cancels the order, sends the invoice or changes the address, depending on which system is connected. The right order for getting started follows at the end of the mistakes section.
Decision guide: which setup suits your shop?
The right level depends on channels, team size and ticket mix. Look for the row in the table where most of the criteria fit your shop:
| Setup | Channels | Team | Fits if | Limit |
|---|---|---|---|---|
| Shared inbox with templates | 1 (email) | 1 person | You are just starting and receive few enquiries a day | From the second channel onwards, you lose the overview |
| Inbox with all channels | 2 to 3 | 2 to 4 people | Instagram or WhatsApp regularly bring in enquiries | A person still answers standard questions |
| Inbox with shop connection, workflows and AI | 4 or more, including marketplaces | 5 people or more, or temporary staff in Q4 | Order status, returns and invoices make up most of the volume | Needs clean order data, a well-maintained knowledge base and lead time before peak season |
The third row is the point where channels and order data have to come together. If your shop is there, we will show you how armincx connects inbox, shop integration and AI in a free demo.
What should you not automate in digital customer service?
Automate what has a clear answer from data, and only open channels your team actually looks after. These cases stay with a person:
- Escalations, goodwill and legal matters: Your team decides on repeat complaints about the same lost parcel, refunds above, say, €100 and statutory warranty questions.
- The first reply to VIP customers: Anyone who has bought from you for years gets a personal message.
- Public Instagram comments: Everyone can read along here. The AI may prepare a draft, but a person should send it.
- Live chat outside service hours: Show a contact form instead, which lands in the inbox as a ticket.
- Phone for products that need explaining: Keep the hotline if customers want advice before they buy.
For these cases, you need clear handover rules. Define the keywords, customer groups and amounts that send a ticket straight to a person.
Which mistakes slow digital customer service down?
When you switch, the order decides speed and quality. These five mistakes hold you back:
- Opening all channels at once. Start with the two channels that lead your count and add more once the inbox is running.
- Launching a chatbot without order data. It points to the contact form for every parcel question, and the customer writes a second message.
- Leaving channels in separate apps. Instagram on the phone, WhatsApp in the browser, email in Outlook: this is how duplicate tickets and contradictory answers happen.
- Writing the knowledge base once and then leaving it. Otherwise new delivery times or return periods reach customers as wrong answers.
- Switching in November. Plan it so your team has used the inbox for a few weeks before peak season.
You avoid all five by following the order of the five-step plan. Our guide to customer service automation has a detailed matrix of which topics suit automated answers and where to start.
FAQ: common questions about digital customer service
What does customer care include?
In a shop, customer care covers all enquiries before and after purchase, from product advice through order status and returns to complaints. In digital customer service, they come together via email, WhatsApp, Instagram, Facebook, live chat and marketplace messages in one shared inbox.
Which channels matter most in digital customer service?
In everyday shop life, email is by far the largest service channel, followed by Instagram and WhatsApp. A two-week count of your own enquiries shows which channels matter for you.
Is email still important in digital customer service?
Yes. Email is the largest service channel in everyday shop life and therefore belongs in every inbox.
Is digital customer service worth it for small shops?
Yes. Even with two channels, a shared inbox with order data saves your team duplicate replies and searching the shop backend.
Does digital customer service replace the phone?
It depends. For products that need explaining, the phone remains an important channel. You answer standard questions about orders and returns faster in writing, because the answer comes straight from the order data.
Do I need AI for digital customer service?
No. The foundation is a shared inbox with a shop connection. AI pays off once many recurring enquiries such as order status or invoices come in.
What does digital customer service cost?
Costs consist of software, setup and your team's time and depend on channels and ticket volume. For WhatsApp, usage-based fees from Meta may apply.
How do I get started with digital customer service?
Count all enquiries by channel and topic for two weeks. Then bring the two largest channels into one inbox and connect your shop.
Conclusion: digital customer service starts with a count
Digital customer service means all channels in one inbox, order data in the ticket and clear rules on what a person decides. The order matters more than the tool. You count first, then bring the channels together, connect your shop and only then automate.
How much runs automatically depends on your setup, which you build in these steps. For the most common enquiry, our solution for Where is my order requests shows what the process looks like with order data.
Your next step: start the count on Monday and after two weeks check which five topics fill your inbox.
