AgentovoFor online shops
Make customers wait three days, and they’ll buy elsewhere.
We’ve run online shops ourselves. One time the inbox buried us. The other time, there was nobody to advise customers online. Those are exactly the two problems we now solve for other shops.
Support inbox
We know this one from our own inbox.
The same questions every morning. Where’s my parcel, can I exchange this, where’s the returns label. None of them is a big deal on its own, but together they eat half a working day. The assistant drafts the replies, you approve them.
InboxMonday, 8:10
- Where’s my order 4821?Draft ready
- Can I swap this for a different size?Draft ready
- I need a new returns labelDraft ready
- Please change my delivery addressDraft ready
- When will the serum be back in stock?Draft ready
- Complaint, photo attachedNeeds you
- 1
We read along.
The assistant learns from your past replies and from your shop and shipping data, not from some generic FAQ. It sounds like you.
- 2
You approve.
Every reply lands in your inbox as a draft. You read, tweak and send. If it doesn’t know something, the question comes to you.
- 3
You let go, one step at a time.
Only once you trust a type of enquiry, like the parcel question, does it go out on its own. Anything sensitive stays with you. You decide, not us.
Online product advice
In your shop, you ask questions. Online, nobody does.
AdviceExample
Which cream is good for dry skin in winter?
Can I ask two quick things: does your skin feel tight, or is it flaky too? And are you fine with fragrance?
Mostly tight. And ideally fragrance-free.
- Rich, fragrance-free night cream – for tightness overnight
- Lightweight hydrating serum – goes under any day cream
- Balm for on the go – for when it’s cold out
Online skin quizzes and product finders are decision trees: three clicks, one result. That works for a brand with two products. It doesn’t work for a specialist retailer stocking lots of brands.
The assistant advises the way you do in your shop. It asks questions, suggests a few products that fit and explains why. Your customer makes the choice.
What a typical chatbot can’t do.
The difference is in the details, the ones you only notice three months in.
| Topic | Typical chatbot | Agentovo |
|---|---|---|
| Answers | from an FAQ | from your own past replies |
| Email inbox | usually not included | the core of it |
| Advice | decision tree, one result | asks questions like in store, several reasoned suggestions |
| Product range | fixed logic | many brands, easy to extend |
| Control | runs however it runs | every reply can be approved first |
| Updates | when the vendor ships one | whenever you correct something |
| Cancellation | everything’s gone | system and workflows stay with you |
What you can add later.
You start with one of the two. What comes next is up to you, once the first is up and running.
- Website and WhatsApp chatUses the same knowledge as your inbox, so customers get the same answer wherever they ask.
- Shop and shipping integrationOrder status, tracking, returns: the assistant looks things up instead of guessing.
- Stock across every channelSell something on Amazon and your shop’s stock level updates. No AI here, on purpose, because stock numbers have to be exact.
- After the saleReview requests, reorder reminders, and a message when something is back in stock.
- Phone assistantWe build these too, but for most shops we’d advise against it: your customers write, they don’t call. We’ll tell you straight whether it’s worth it.
- Chat and email assistants, connected to the shop and shipping.
- A dashboard for management: every question live, upload your own knowledge, time saved at a glance.
Your system, your data.
In FrankfurtThe AI models run on Google Cloud in Frankfurt. Your data stays in the EU.
Your own accountsDatabase, automation and voice run on accounts set up just for your shop, never shared.
Yours to keepIf you cancel, you keep the system. We don’t take anything away.
Not ready yet?
The honest e-commerce guide.
24 pages on the ten places where online shops lose money. For each one, we tell you straight whether you can fix it yourself, need a specialist, or whether it’s something we do. Currently available in German only.
- A self-check: twelve questions, five minutes
- The maths on when automation pays off
- The 2026 legal landscape in Germany, including the new withdrawal button
- Quick wins you can put in place this week, no agency needed
What shops ask us.
When does it pay off?
As a rough guide, from around €15,000 in monthly revenue, or once enquiries take up an hour or more of your day. With only a handful of orders, a good set of reply templates is the better investment. We’ll tell you in the first conversation.
Do we need to write a brief?
No. For the inbox, we read your past replies. For product advice, you advise us once, as if we were a customer in your shop. We build the logic from that.
Will it work with our shop system?
We connect to the inbox, shop system and carriers you already use. If your shop system is small or unusual, the advice runs on its own page under your domain. No migration, no plugin.
What about complaints and tricky cases?
They stay with you. The assistant spots them and passes them to you instead of replying itself.
How long does setup take?
A few weeks. After that, everything runs with your approval first, for as long as you like.
Where is the data stored?
Every shop gets its own accounts for database, automation and voice. The AI models run on Google Cloud in Frankfurt. Your data stays in the EU.
Will it pay off for your shop?
The free Readiness Check tells you in 30 minutes. You come away with a playbook and clear next steps, whether or not we work together.
Book your free AI Readiness Check