Services/AI Consulting

ai
consulting

We start with what eats time in your business every day, not with the technology. For about half of it the answer afterwards is: this needs no AI.

LocationVienna, for all of Austria
Entry pointAssessment
After thatImplementation if it pays off
What we do not doTraining, funding advice
Vol. 17 // No. 2026Ed. KI

what ai consulting
means with us

AI consulting with us means three things, in this order: looking at where there is anything to gain in your business at all, then building what pays off, and saying so when it works better without AI. We come from development, not from management consulting. That means the result is something that runs, not a presentation about something that could run.

look first, build second

STANDORTBESTIMMUNG

We go through with you which tasks come up daily, how often and how long they take. Out of that comes a list of what pays off and what does not. For most companies that is two or three items, not twenty. The result is a document you can also take to somebody else.

we build it too

EINFÜHRUNG // UMSETZUNG

The difference to pure consulting: what tops the assessment we can then build. A chatbot for recurring customer questions, reading invoices and receipts, product texts in several languages, search across your own documents. That is development work, and we have done it since 2018.

often no ai is needed at all

AUTOMATISIERUNG

A good share of what companies describe as an AI problem is a missing interface. Quotes typed by hand into the second system. Appointments maintained twice. That can be solved without AI, more cheaply and more reliably, and we say so instead of selling you a model.

what ai is suited for,
and what not

This is the content of an assessment, condensed into a table. For two of the five task types we advise against it, and not because of the technology but because accountability stays with you and you have to be able to explain every decision, including in two years.

What AI is suitable for and what not
Always the same, many casesReading invoices, matching receipts, syncing data between two systemsworth itThis is where automation pays off fastest, and often it needs no AI at all, just a clean interface. That is cheaper and more reliable.
Text in, text outAnswering recurring customer questions, product texts in several languages, summaries of long documentsworth itThe area where current models are genuinely strong. The condition is that a human looks at the result before it goes out.
Searching your own documentsFinding answers in manuals, contracts or old quotesdependsWorks well when the documents are organised and current. With a grown folder mess you automate the mess along with it, and then the answer is worse than before.
Decisions with consequencesAssessing creditworthiness, pre-sorting applications, setting pricesbetter notNot because of the technology but because of accountability. You have to be able to explain every one of those decisions, including in two years, and with a model that only works to a limited extent.
Rare but importantOne special case a month that looks different every timebetter notThe setup effort bears no relation to twelve cases a year. Tasks like that are better left with the person who does them best anyway.

Qualitative assessment from our practice, without absolute values

What we are actually talking about

four things we
actually build

With AI there is a lot of talk about possibilities and little about results. So here are four applications we build, along with the condition under which they pay off. If your case is not among them that does not mean no, but we look more closely then.

Automating customer service

A chatbot that answers the twenty questions arriving every week and hands everything else to a person. The gain is not replacing staff, it is that nobody types the same answer for the hundredth time.

Reading invoices and receipts

Capturing amounts, dates and suppliers automatically and handing them to the accounting system. One of the few tasks where the setup pays for itself in the first year, provided enough receipts come in.

Producing quotes faster

Generating a draft from text blocks and past quotes that a person checks and sends. Saves several hours a week at companies with many similar quotes.

Search across your own documents

Finding answers in manuals, contracts and past projects without knowing the file. Requires the documents to be reasonably organised, otherwise you automate the mess along with it.

Bernhard Müllauer, founder of creatingfaces and your contact for the project
Your contact

bernhard müllauer

I write the quote, I design, I build. You do not get an account manager standing between you and the work, you talk to the same person from the first question until after launch.

If something does not make sense, I say so, even when that makes the project smaller. In the long run that is more pleasant for both sides. The first conversation is about your processes, not about tools. It takes an hour and costs nothing.

why there are
no prices here

On every other page of this website there are numbers, and we think that is right. Not here, and there is a reason: the scope varies too much. Two companies with the same question can end up at one day or at three months, depending on what their data looks like and how many systems are involved. A number that does not fit half the cases is not information, it is a trap.

What you get instead: a conversation at no cost, then a fixed price for a clearly defined first step. No hourly rate, no open-ended invoice. If after the conversation it is clear that it does not pay off, we say so and send no invoice.

what we do
not offer

No training and no seminars. There are good providers for that, we are not one of them. No funding advice: there is funding for digitalisation projects in Austria, but we give no advice on it you could rely on. What we do provide is a clean breakdown of the project you can take to a consultation.

If you sit in Vienna and would rather start in person, more on that is on our page about the AI agency in Vienna. If it is an application with users and a database rather than consulting, then web app development.

what eats your
time

Describe one task that recurs every week in your business. You get an honest assessment, even when that assessment is that nothing here pays off.

[ More Services ]All Services →

questions & answers

[ FAQ ]

We deliberately name no price on this page, and it is the only page of our website without numbers. The reason: the scope varies too much. Two companies with the same question can end up at one day or at three months, depending on what their data looks like and how many systems are involved. You get a free first conversation and then a fixed price for a clearly defined first step, not an open hourly rate.

With an assessment. We go through which tasks come up daily, how often and how long they take. Out of that comes a list of what pays off and what does not. For most companies that is two or three items. The result is a document you can also take to somebody else.

Well suited are recurring tasks with many similar cases, and anything where text goes in and text comes out, such as recurring customer questions or product texts in several languages. We would advise against decisions with consequences like creditworthiness or pre-sorting applications, because you have to be able to explain every one of them, and against rare special cases where the setup never pays off.

No. There are good providers for AI training and seminars, we are not one of them. We come from development, meaning we build things that run afterwards. If you are looking for training we say so right away instead of offering one we are not good at.

No. There is funding for digitalisation projects in Austria, but we are not funding consultants and give no advice on it you could rely on. What we do provide is a clean breakdown of the project with scope and effort that you can take to a consultation.