Since July 2026, there's been a new face on our team page. Her name is Anna Stickler, she answers support requests around the clock, and she hasn't taken a single sick day. The reason is simple: Anna isn't human. She's our first AI team member.
And that's exactly how it's stated on her profile. Right next to her name sits a clearly visible badge: "AI Team Member." No asterisk, no fine print, no hidden footnote tucked away at the bottom of the page. Anyone chatting with Anna knows exactly who they're talking to.
Why We Chose an AI Team Member
If you sell software, you also sell support. Over the years, we've developed more than 120 Shopware plugins, around ten of which are currently listed in the Shopware Community Store, rated 4.9 stars. This wealth of experience shows us exactly where the pain points are. Every extension brings its own set of questions: Why isn't the invoice showing up in the customer account? How do I set up X-Rechnung correctly? Why did the integration stop working after the Shopware update?
These questions are legitimate. And they don't come in between nine and five. They come on Sunday evenings, when a shop owner is restructuring their product range. They come at two in the morning, when someone's sitting in a different time zone. They come in droves, right after a new Shopware release drops.
Our team is good. But it's a human team with human limits. In the past, a ticket that came in on Friday at 6 p.m. often didn't get a reply until Monday. For a shop owner whose invoicing is broken over the weekend, that can feel like an eternity.
We had two options. Hire more people who could also be reached at night and on weekends. Or find a different answer. We chose the different answer. And since we build an AI helpdesk platform ourselves, that answer was closer at hand than you might think.
Who Anna Is and What She Does All Day
Since mid-July 2026, Anna has independently handled first- and second-level support for our Shopware plugins. Fully. In concrete terms, that means:
A customer submits a support request through the Shopware Store, by email, or via our web form. Anna reads the request, understands the problem, and taps into our knowledge base. She knows the documentation for every plugin, she knows the typical pitfalls, and she knows the solutions from thousands of past tickets. Then she responds — usually within minutes, no matter the time of day.
Here's an example from the first few weeks: a customer using our PDF and e-invoice plugin reported that automatic ZUGFeRD invoices weren't being sent to his customers. Anna interpreted the screenshot of the error message, identified the cause (a misconfiguration in the mail delivery settings), and sent the customer a step-by-step guide. Problem solved — without a human ever laying eyes on the ticket. At 10:47 p.m.
That's become the norm. Anna works through tickets, answers questions, and documents solutions. And the truly tricky cases, she hands off to us. More on that in a moment, because that's exactly the part we're most proud of.
Transparency Isn't Just for Show
Plenty of companies use AI in customer support and stay quiet about it. The friendly "Sarah from the support team" who's actually a chatbot. We think that's a mistake.
Sooner or later, customers notice whether they're talking to a machine or a person. And once they realize they've been misled, trust is gone — not just in the chatbot, but in the entire company.
That's why Anna is listed openly on our team page. With a badge. Anyone who gets a reply from her knows an AI answered. That sets a different kind of expectation, and it's healthier for both sides. The customer knows they can get a human if they need one. And we don't have to play a role that benefits no one.
This openness wasn't a marketing decision for us. It's part of our conviction about what AI in a business should look like: visible, traceable, and controllable.
The Technology Behind Anna: inbrix.ai
Anna doesn't run on some off-the-shelf solution. She runs on inbrix.ai, our own AI helpdesk platform. In other words: we use the product we built for other support teams for our own support as well. We eat our own dog food, as they say in the software world.
That has a practical side effect. Every weakness we notice in our day-to-day work with Anna flows straight back into development. Anna isn't just our support team member — she's also our toughest tester.
Your Tickets Never Leave the EU
Let's start with what's the top criterion for many shop owners in the DACH region: data protection. inbrix.ai is fully GDPR-compliant, and hosting takes place in Germany, or the EU more broadly. Our guiding principle is: your tickets never leave the EU.
That might sound like a given, but in the world of AI, it isn't. Many helpdesk tools route support requests through American data centers and AI models whose data processing lands in a gray area. For a support ticket containing an order number, an email address, and maybe a screenshot with customer data, that's a real problem. We built things differently from day one.
Every Channel, One Inbox
Support requests don't come through just one channel. One customer sends an email, another uses live chat on the website, a third submits a ticket through the Shopware Store. In the past, that meant three systems, three inboxes, and three times the risk that something would slip through the cracks.
inbrix brings all of that together in a unified inbox. Email, chat, web tickets, and Shopware tickets all land in one place. Anna sees everything in a single stream and can handle every channel the same way. For us, that means no request falls through anymore, no matter how it comes in.
Two Checks Before Every Single Reply
Now for the part that sets Anna apart from a simple chatbot. Before Anna sends even a single reply, she runs through two checks.
The first is a competence check. Essentially, Anna asks herself: Can I answer this safely and correctly? Do I have the information I need in the knowledge base? Is my answer well-founded, or am I just guessing? Only if this check comes back positive does the process continue.
The second is a risk assessment. Anna gauges how sensitive the topic is. Is this a harmless configuration question, or something where a wrong answer could cause real damage? Only once both checks give the green light does Anna respond on her own.
These two guardrails are the reason we trust Anna in the first place. An AI that simply answers everything, no matter what, would be a liability. An AI that checks whether it should even be answering before it does is a tool you can actually work with.
Where Anna Stops and We Take Over
There are certain topics Anna always hands off to the human team, without even attempting them first. These include:
- Pricing questions and quotes. Talking about money is something we humans do. Full stop.
- Legal questions. If a customer asks about liability, contracts, or licensing terms, that's not a job for AI.
- Login credentials and security-related matters. Anything involving passwords, access, or sensitive configurations goes straight to a human.
The handoff itself runs smoothly. Anna passes on the ticket along with the full conversation history, so the colleague picking it up is immediately in the loop and doesn't have to start from scratch. The customer doesn't experience an abrupt break — just a handoff that works.
Anna Gets a Little Better Every Day
inbrix learns from resolved cases. Every ticket a human handles, every solution that gets documented, makes Anna more capable the next time a similar case comes up. This isn't a static rulebook someone wrote once and left unchanged forever. It's a system that grows alongside the work.
Anna can also interpret screenshots, which is enormously helpful in technical support. A customer who snaps a photo of an error message doesn't have to painstakingly type it out. Anna reads it and understands the context.
And because at the end of the day you need numbers, inbrix tracks two metrics that matter to us: cost per ticket and customer satisfaction via CSAT score. That way, we can see in black and white whether Anna isn't just faster, but actually good.
How the First Few Weeks Really Went
Honestly, we were nervous at first. Support is a matter of trust, and we handed over a large chunk of it to an AI. What if Anna said something wrong? What if customers felt like they were being brushed off?
None of that happened — at least not to the extent we'd feared. Response times dropped dramatically. Anna typically replies within about five minutes, around the clock, whereas a human, depending on workload, sometimes needed hours. A ticket from Friday evening used to risk sitting untouched until Monday. The CSAT score for tickets handled by Anna is on par with our human team's — and for some standard cases, even higher, simply because the answer arrives faster.
What surprised us: customers respond positively to the transparency. Instead of feeling put off that an AI is answering, many appreciate getting a competent answer right away. And anyone who genuinely wants a human just says so — and gets one.
There were, of course, a few bumps along the way. In the early days, Anna escalated a handful of cases she could actually have resolved herself, because her competence check was set a bit too cautiously. Better that way round than the reverse. We fine-tuned the thresholds, and by now Anna makes that call much more reliably.
But the biggest effect shows up in the people on the team. They no longer spend their time answering the same configuration question for the fiftieth time. They focus on the tricky cases, individual consulting, and improving the plugins — the work that actually needs a human touch. Anna takes the routine off their plate, not the job itself.
What This Means for Our Plugin Customers
For everyone using our Shopware plugins, support has noticeably improved. Faster answers, around the clock, with consistent quality. The common questions we're familiar with from everyday e-commerce operations get answered by Anna in seconds.
At the same time, no one is left at the mercy of a machine. When things get complicated, when it's about money or legal matters, or when a customer simply prefers to talk to a human, that's always an option. We deployed Anna to make our support better, not to hide behind her.
What You Can Take Away for Your Own Support Team
You might recognize your own situation in our story. A support inbox that never empties out. Requests that pile up outside business hours. A team answering the same questions over and over instead of working on what really matters.
From our experience with Anna, a few takeaways apply to any support team:
Transparency pays off. Being open about using AI builds far more trust than concealment ever could. The badge on our team page isn't a risk — it's an advantage.
Guardrails are a must, not a nice-to-have. An AI without a competence check and risk assessment is a ticking time bomb. The value of AI in support doesn't lie in answering everything — it lies in knowing when to stay quiet and hand off to a human instead.
Data protection isn't a formality. Especially in the DACH region, the question of where data is stored isn't a nice-to-have. Support tickets contain customer data, and that data has no business being routed through American data centers.
AI doesn't replace people — it frees them up. The biggest win for us wasn't cost savings. It was the redistribution of work. The people on our team are now doing the work they were actually hired to do.
Anna Is Just the Beginning
With Anna, we've shown what's possible with AI in support when it's set up the right way: transparent, controlled, and GDPR-compliant. And the technology behind her, inbrix.ai, isn't just available to us.
If you're running a support team that's buckling under the weight of incoming requests, if you want to deliver faster answers without sacrificing quality, and if it matters to you that your customers' data stays within the EU, then inbrix is exactly what Anna does for us — just for your team.
Anna's working right now. There's a good chance she's answering a ticket at this very moment, while you're reading this. Maybe it's time your support team got someone like her, too.
Check out inbrix.ai and find out what an AI team member could look like for your support team.