For companies and teams
“We have to do something with AI. I just don’t know where, and I don’t have the time.” I hear that sentence from managing directors almost word for word.
We build the two or three things that noticeably give your team time back. Afterwards your people run them, not me.
I’m Mario Fraune, AI sparring partner based in Vienna.
How do you work?
What I see in companies
If a process somehow worked before, it will somehow work faster afterwards. It is not the technology that is to blame, it is the step that was skipped before it: rethinking your own system with the possibilities of AI, instead of speeding up the old paths.
That is normal: whoever hands out tasks loses sight of details that everyone used to know. An example from my work: someone has been working for months without the paid version of their most important tool, while elsewhere a new system is being considered.
The request is: we need AI for this and that. But the time rarely goes into the first result, it goes into what happens to it afterwards: the follow-up, the handover, the correction loop. Two hours a day into emails, because there is no template for the cases that keep repeating.
Sometimes it is true. Often nobody has looked in years to see whether it still holds today. An old caution has turned into a rule, and the rule blocks a topic that has long been solvable.
What ends up running at your company fits your business and no other. That is why I look first at where someone waits for someone else and what holds your people up every day.
Then I ask the questions that get something moving. Sometimes the answer is an AI system, sometimes it is a template that has been missing for years. I will tell you both.
The benchmark is a system nobody talks about after four weeks, because it simply runs. Slide battles and a recommendation after two meetings are available elsewhere.
“Good design is actually a lot harder to notice than poor design, in part because good designs fit our needs so well that the design is invisible, serving us without drawing attention to itself.”
Don Norman, author of “The Design of Everyday Things”
Typical starting points from my work, each with what we built and what was different afterwards.
Starting point Intervention Result
Team · eight people
Agency · twelve people
SME · thirty people
Service firm · twenty people
Engineering office · ten people
Your result is missing
You tell me how work happens at your company and where it gets stuck. At the end you have an assessment of where your first sensible step is. Whether you take it with me is your decision afterwards.
What your team puts in 30 minutes, once Just you, nobody else.
I look at how work really happens at your company: where somebody waits for somebody else, what holds your people up every day. Out of that comes a sorted list of what pays off first, often a shorter one than expected.
What your team puts in One conversation per area involved The people who run the process day to day.
We take one core process and set it up again with AI: built together, documented, your team trained on it. Your people are in the room instead of being handed a finished system at the end.
What your team puts in About two weeks of guidance The same people as in step 02.
What is built goes into the team: practised on real tasks from your company, not on examples from a template. Accounts, data processing agreements and a few rules of thumb end up on one page.
What your team puts in Half a day or a full day Everyone who will work with it afterwards.
The process runs, your team maintains it itself, and there is a list of what comes next. If you want me to stay in the background, you keep me as your external AI lead: monthly, and cancellable monthly.
What your team puts in No fixed effort after that The process belongs to you.
Where which data ends up is part of what decides which solution is even an option. That is why the question comes at the beginning with me, not at the end. If you have a data protection officer at your company, I work in sparring with them: they review, I build, and we clarify up front what is possible instead of finding out afterwards what was not. In my experience the room to move is bigger than assumed internally.
Less than the headlines suggest. Most companies are users (“deployers”), not providers of AI systems. Three things matter above all: AI literacy in the team, the transparency obligations in force since August 2026, and a short look at the risk class of the systems in use.
In the free version: no. The rule of thumb is “case, not person”: describe the situation, not the human behind it. For company data you need at least a paid plan, a data processing agreement with the provider, and AI training switched off.
No. Free versions often use your input for training and offer no data processing agreement. The minimum for business use: paid plan, processing agreement, training opt-out. With the major providers that is set up in a few minutes.
Yes, on several levels: EU providers such as Mistral with Le Chat, platforms with EU hosting such as Langdock, or fully local models on your own network. Often the US provider with the right settings and contracts is enough as well. What fits depends on the kind of data you handle.
Which one fits you is something we clarify in the first conversation. I will not commit before that, because I do not yet know where your lever sits.
Entry
Your team works on real tasks from your company, not on examples from a template. I prepare the cases with you beforehand. As a half-day or full-day format, remote or on site with you.
One process
We take one core process and set it up again with AI: built together, documented, your team trained on it. At the end there is a before-and-after look at the time the process used to cost and the time it costs now.
Long-term
Your external AI lead, without creating a position: strategy, tool decisions, questions from the team, implementation. Monthly, and cancellable monthly.
What a format costs is something I tell you after the first conversation: one price, everything included, preparation and travel as well. No hourly lists, no add-ons.
From my own experience
Your team decides whether it works. So I bring your people in early instead of putting a finished system in front of them.
People who resist a change almost always have good reasons: a grown process holds years of work, and whoever runs it every day knows every exception that breaks an automated procedure. So I ask for advice before I touch anything.
“Let’s look together at what works better
and where AI can do more for you.”
The supposed blocker becomes your most important reviewer that way. That is not politeness, it is the reason the solution actually gets used afterwards.
The car is a toy. The system behind it is not.
For more than fifteen years I built digital products, most of that time as a UX designer. Back then many people thought it was about making interfaces prettier.
At its core it was always about understanding what someone needs at a given moment. Which means asking a lot of questions and challenging existing systems, especially the ones everyone takes as given.
With AI it is the same, only with higher stakes. Anyone who simply makes the old paths faster misses the actual point.
“You have already heard the standard answer 30 times.
I am the call after that.”
It started with a documentary about a game of Go: the machine makes a move nobody in the room understands, and the player later calls it beautiful.
Calm and consistent towards one goal: that is how I have gone into every project since. We want to win by a mile, and that wish narrows the strategy before the first move is made. Half a point is enough for the machine, so it calmly plays forty moves ahead.
The documentary is called AlphaGo and is on YouTube for free.
For a year and a half I have been building a system that carries the larger part of my own business. I do not sell you anything I do not use myself every day.
Home is Vienna, the work happens remotely across the German-speaking region. For workshops I come to you.
I keep trying out whatever is new and watch which of it holds up. So you get the solution that fits your process, not the one everybody happens to be talking about.
Nobody at your company has to write a line of code for it. Anyone who wants to know what is going on under the hood can look in with me. We do that together.
For fifteen years it was my profession to watch people work and understand why a tool ends up unused. So I ask first where the time burns, and only then which technology is a candidate.
I have built and led a team myself, from newcomers to colleagues with decades of experience. That is not something I learned in a seminar.
Whether we are a fit is something we find out in 30 minutes.
From workshops, webinars and calls. Answered briefly.
Almost always the same two patterns. First: speeding up routine texts, meaning proposals, emails, minutes. 45 minutes become 8. Second: making your own knowledge searchable, meaning contracts, projects and instructions as a queryable knowledge base. Unsexy, but exactly these two make the difference for my clients.
Because otherwise you end up with a system nobody at your company understands, and a service provider you will never get rid of. If we build it together, both belong to you afterwards: the process and the ability to build the next one yourselves.
Remote is the default, and it works excellently for sparring and workshops. For team workshops I also come on site, anywhere in the DACH region. Travel costs are part of the flat price, not a separate invoice.
I will tell you after the first conversation, as a flat price: one price, everything included, preparation and travel as well. No hourly lists, no add-ons. Any number before that would be guesswork, because I do not yet know where your lever sits.
How we start
The compass call, 30 minutes, free. I ask you the questions that probably nobody has asked you on this topic before.
Afterwards you know where to start. Whether you continue with me is your decision, in your own time. It is not a sales pitch, and there are no slides.
30 minutes, free, no sales pitch.
