You stop typing the same data in again and again, and stop digging through folders months later. I connect the programs you already use so they hand the data to each other by themselves — and your company data stays with you.
Based near Ingolstadt/Regensburg. I travel on site for the rollout; ongoing support runs remotely. I work in German and Polish, and correspond in English.
A practitioner, not a salesperson: what I offer, I run myself every day — at an e-commerce company and on my own servers.
own servers in daily use for yearsIT diploma recognised in Germany (IHK FOSA)
Services
I do three things, and you can order each one on its own:
→a computer at your premises, or a rented machine used only by you — your company's data and programs stay there;
→your documents in one place, with a chat that searches them and shows where the answer came from;
→a link between the programs you already use, so they pass data to each other by themselves.
We only start with the computer if the one you have now is not enough, or the data cannot leave the company.
Your company data stays with you
problem
Your important paperwork sits in a handful of online services whose rules somebody else sets — and every month brings another subscription. Every AI idea runs into the same question: are we even allowed to upload our documents there?
what I do
I set up and run a company server built on tools: I keep an eye on the running services, make backups, and test by restoring that they actually work. Your data stays where you have access to it and know who can see it.
in practice
My own infrastructure runs on the same principle, documented publicly on tetes.pl — with the services watched, backups made, and checks that they can actually be restored.
Instructions, procedures and answers to common questions are scattered across PDFs, emails and notes. Nobody knows which version is current.
what I do
I build an assistant that finds the right passage and answers in plain language, showing its source. Most of the work is putting the documents in order, not the AI itself — I say so upfront, because that is what decides whether the answers are worth anything.
in practice
The demo below works exactly this way — ask it about my articles and projects. It is designed to answer "I did not find that" rather than guess when nothing matches. That is a deliberate choice: a company is answerable for what it shows its customers, including the words its assistant produced.
Invoices, letters and scans end up in folders, e-mails or binders. After a few months, finding one specific document becomes difficult.
implementation
A digital archive recognises the text on scans and photos, so you search by content, date or kind of document — not by file name. I configure tags and assignments to match the company's documents; over time the system applies them ever more accurately on its own.
result
A document can be found with a few words instead of manually digging through folders and attachments.
The same data has to be moved between a shop, e-mail, a spreadsheet and invoicing. Every manual step adds delay or risk of error.
implementation
Automation connects existing tools and performs repeatable steps by itself, notifying a person when something fails.
result
831 times in seven days the process ran by itself, with nobody having to stand over it (measured 22 Aug 2026).
What does this look like in a real business?
Open the trade closest to yours. I show step by step what the system does — using tools I work with every day. These are not accounts of client projects, but workflows that can be set up at your company. What you gain from them depends on your business.
Car workshopInquiries from three channels on one list — and a reminder when a customer's service is due again.
how it works today
Inquiries arrive by phone, by email and through Facebook. Some end up on a scrap of paper, some stay in your head. The customer from a year ago does not come back, because nobody reminded them.
what the system does
1Emails and Facebook messages land on one list by themselves; calls you add with one click or dictate on the road.
2If an inquiry sits unanswered longer than you decide, the system tells you — not the customer.
3Each vehicle carries the date of the last job. A year after the oil change, or before the roadworthiness test, the customer gets a short message with a suggested slot.
what you get out of it
Inquiries stop sitting in three places at once, and some customers come back on their own, because the reminder arrived in the right week.
Online shopPrices and stock at suppliers and competitors checked every morning — one message instead of thirty browser tabs.
how it works today
A supplier quietly raises a price or drops a product. You find out when a customer has ordered something that no longer exists, or when the margin has gone to zero.
what the system does
1Every morning the system works through the products you name and records prices and availability — at suppliers and wherever competitors sell.
2It compares against yesterday and filters out small fluctuations itself, so you are not buried in noise.
3When something really moves, you get one message: what, where and by how much.
what you get out of it
You usually learn about a raised price, or a product the supplier has dropped, in the morning — for the products you flagged to watch.
Construction companyA photo and two spoken sentences on site become a dated entry attached to the job.
how it works today
The paperwork gets done in the evening, after work, from memory and from several hundred photos on a phone. If a complaint comes a year later, nobody knows which photo is from which day.
what the system does
1On site you take a photo and speak two sentences into your phone.
2The system recognises the speech, adds the date and the location from the phone, and attaches the entry to the specific job.
3Everything from one job sits together and can be searched by content, not by file name.
what you get out of it
The entry is made on site rather than in the evening from memory — and in a dispute over a defect you reach for a dated record.
Practice or officeYou ask in plain language and get the answer along with the document it came from.
how it works today
One person knows the procedure. When they are away, the others guess, or call the client to ask something the company settled long ago.
what the system does
1Referrals, results, contracts and instructions go into one place — including scans and photos, because the text in them is read out.
2Each document shows when it arrived and which version is the latest. Older ones stay, but are marked as no longer in force.
3You ask in plain language: “from when does this price list apply?” — and get the answer with a link to the document it came from.
what you get out of it
A new colleague does not have to ask anyone, and sees which version the answer came from.
Restaurant or caféA new review reaches you straight away, with a ready draft of a friendly reply.
how it works today
You notice reviews from Google and other sites a week later, and by then the reply looks forced. A new guest reads the last five, and those decide whether they book a table.
what the system does
1A new review from Google and other sites reaches you within minutes of being posted.
2A draft reply comes with it — to read and adjust, not to send blind.
3The review keeps a record of who replied and when, so nothing goes unanswered by accident.
what you get out of it
You reply the same day, while it still counts — instead of noticing the comment a week later.
Live demo: an AI assistant answers questions about my projects and articles — and cites its sources.
An AI model (Cloudflare Workers AI) reads your question only to write the answer. Nothing is kept afterwards. Please do not type in personal data.
The same assistant can answer questions about your company's documents.
How I work
1
A short talk about specifics
I ask about a few real things from your business: where time goes, where paperwork disappears, what repeats every week. 30–45 minutes, remote, free. After the call you get 1–3 ideas in writing on where to start — or an honest “there is nothing here worth automating”, if that is how it turns out.
2
A small plan, no slides
Then I propose one or two small things to try. For each I write what I will do, by when and for how much — one price, no running hourly meter and no fifty-slide deck. You have probably heard the stories about systems bought sight unseen: countless features nobody uses and no way to adjust anything by hand. That is why I start small — with a scope you can check, and reject.
3
Implementation
Remote or on-site — travel within the Ingolstadt/Regensburg region is included in the price.
4
Support or handover
At the end you choose: either I keep running it for a fixed monthly figure, or I hand you the whole thing with its documentation and you do not depend on me. The prices for both are below.
Support after the rollout — optional, two tiers
In both tiers, outside the hours
Updates, regular checks that the services are running, backups and an attempt to restore the data from them. That runs by itself or I do it — and it does not eat into your hours.
The hours are my time for your requests
€149/month
half an hour a month for changes and questions. I reply within two working days.
€349/month
two hours a month, a reply within one working day, a short monthly report and an agreed window for planned changes.
Bigger work beyond that is quoted separately — I give the price before I start. You do not have to take support at all: I can hand you the system with its documentation, and whoever you like runs it.
Bavaria runs the Digitalbonus programme in two variants: Standard with a lower cap and Plus up to €30,000 for particularly innovative projects; it runs until the end of 2027. Three things are worth knowing upfront — applications go exclusively through the ELSTER business account, only projects that have not yet started are eligible, and not every purchase qualifies. Funds are released in monthly quotas, so applications are periodically closed.
Questions & answers
Short questions, concrete answers — in the order clients actually ask them.
What does it cost? Do I need expensive hardware?
I will not name the price of the work upfront, because it depends on what it has to do — I name it after we talk, and then it is one fixed figure with no hourly meter. The hardware is a separate, one-off matter: a mini-PC without a graphics card is usually enough to start, so a few hundred euros, and it stays with you. More documents or heavier AI models need a stronger machine — I say so before the purchase, not after. Monthly support is optional and costs €149 or €349; what it covers is set out below.
Is my data safe? Where does all this run?
By default the solution runs on the company's server or in a private EU cloud. If data must not leave the company, I use local AI models that work without sending content to an external service. External APIs only when the client consciously agrees and the data type allows it. If I process personal data on your behalf, we put the necessary arrangements in writing. I describe where the data physically sits and which services are involved; the legal assessment stays with your advisers.
Does adopting AI bring new obligations for me?
Marketing material often talks about a “training obligation from August 2026”. That is misleading. Article 4 of the EU AI Act has applied since 2 February 2025. Since 27 July 2026 it requires measures to support the development of AI literacy among the people who work with such systems on the company's behalf; it does not require guaranteeing a particular level of competence for each individual. The regulation neither prescribes a specific course nor says that one training session is always enough.
During the rollout I document the technical briefing and the limits of the system. Which further measures your company needs is a question for your legal or data protection advisers.
E-invoicing — does it affect me, and what do I have to do?
It depends on whether you issue invoices or mostly receive them.
If you run a shop and invoice businesses — your invoices may have to go out in a format the other side can read automatically. Whether that applies to you, and from when, depends on the country, the scale and the type of sales — settle that with your bookkeeping. Technically a PDF does not do it, however neat it looks. This usually means real work inside the shop itself, and that is my ground: I have worked with shops for years.
If you mostly receive invoices and a bookkeeping office handles the accounts — your part is small. It comes down to not deleting the data file — that file is the actual invoice, not the printout. Sometimes it arrives as a separate XML file, sometimes it sits inside the PDF, so a neat-looking PDF tells you nothing about whether the invoice can be read by a machine. The office takes care of the rest, and usually does it well.
Deadlines and thresholds depend on turnover and the type of business, and they keep changing over the next few years. What applies to your company is for your tax adviser to determine — I advise on neither tax nor law.
Will my employees manage it?
Whether the whole thing works depends on it, so I treat it as part of the job rather than an extra. The team gets a screen that looks like a normal app — a box for the question, a form or a chat. Not a panel for specialists. I walk people through it unhurried and leave a short guide in German or Polish, included in the price. The technical side I take on myself, within the hours of the package you choose; if more comes up, I give you the price before I do anything.
What do the first two weeks of working together look like?
In the first week we talk, and you get 1–3 ideas in writing — or an honest “there is nothing here to do”. In the second I show the first working version live, provided I get the logins and the data is in usable shape. I do not promise a date before I have seen that; if something drags on, I say so straight away. Before you approve the paid part you can walk away with no obligations.
How long does an implementation take?
I give a date once I have seen the data, the access and the programs that are meant to connect — before that it would be guesswork. What I can say upfront: with an AI assistant most of the work goes into tidying the documents, not the AI itself. I always start with a small scope that shows quickly in practice, and only then expand.
What does implementing an AI assistant on company documents look like?
First we organise the documents: instructions, PDFs, system exports, emails or an internal wiki. Then I build an index so the system can find the right fragment even when the question uses different words than the document. Only then does the AI model write an answer and show the sources it used. Technically this is called , but for the user it should feel simple: ask a question, get an answer, check the source.
How is search different from an AI assistant?
Search shows matching documents — like an internal Google that understands the meaning of the question, not just keywords. An AI assistant goes one step further: it reads the found fragments and writes a short answer. Starting with search is often useful because it quickly shows whether the documents are organised well enough.
What does process automation look like?
I start with today's process: where the data comes from, who copies it, who waits for the result. Then I build a flow that connects existing systems and performs repeatable steps automatically. Error handling matters: if something fails, the system should notify a person instead of silently stopping. At the end you get documentation and handover.
Do I need to know these technologies?
You do not have to. You should get something that works, a description of how it works, and a clear statement of who looks after what — the tool names you can safely skip. I pick them to fit the problem, not the other way round, and I take ones with no licence to pay for. If somebody ever takes this over from me, the documentation holds the full list: , among others , , , and .
About me
My name is Mariusz Binczyk. I administer the systems of an e-commerce company in Germany: servers and email, backups including restores, the network, support for the staff. Plus shops on Shopware 6 and an internal AI assistant that answers questions from the company's documents — the same kind of system I describe above.
I have been building shops and websites since 2008. My qualification is recognised in Germany through IHK FOSA, and I have run my own business since 2017. Privately I run my own servers (): around 50 services that I check on daily, with backups and AI models that run on the machine itself. I write all of it up in public.
I work in German and Polish; documentation and written communication also in English.
