Being a CAIO in Austria

I have been Chief AI Officer at SlopeLift, a performance marketing agency in Vienna, since October 2025. No 500 people, no in-house AI lab, no budget for basic research. What I do have: 18 years of industry experience, clients such as BAWAG, dm and Austrian Airlines, a team that works with Google’s and Meta’s algorithms every day, and the job of integrating AI into real processes before the competition does.

The honest description of my role has little to do with technology. It has a lot to do with asking uncomfortable questions.

The wrong question always comes first

When a company starts using AI seriously, the first question is usually: “Which tool do we need?” That is understandable. And almost always the wrong question.

A CAIO is responsible for making sure AI does not stay a side project and that decisions get made, even uncomfortable ones: Which processes disappear? Which skills become obsolete?

The first thing I did systematically at SlopeLift was a kind of inventory, only the other way round. Instead of asking what AI can do, we asked: What do we no longer want to do ourselves? Which tasks eat up the time of qualified people who need that time for something more important? Where does missing knowledge block the process? Where do we regularly fail at the blank page?

Out of that inventory comes what we internally call the anti-to-do list: a list of tasks that get handed over to agents before anyone thinks about new tools. Sounds unspectacular. But it changes how an organisation thinks about AI: process before tool.

Three categories show up almost every time. First, work nobody wants to do: data cleaning, formatting, repetitive reporting steps. The things qualified people feel are beneath them, but that still have to get done. Second, knowledge gaps: quick research, competitive analysis, briefing preparation. Third, blank-page blocks: first drafts, concept scaffolds, the jump from zero to something.

What the CAIO title really means

The role of the Chief AI Officer in Austria emerges from a concrete necessity: AI projects need decision-making power and their own budget. They need someone who knows the operational reality of the company and at the same time understands where the technology is heading.

In most companies I observe, AI still runs as a pilot project in one department, as a personal hobbyhorse of one executive, as a budget add-on to some digitalisation programme. That does not scale. It fails quietly, without anyone measuring it.

What 18 years of performance marketing have taught

In 2008, 40–60% of my working day was data yoga: export the CSV, build a VLOOKUP, set up a pivot, hunt for errors, replace umlauts, start over. Anyone doing serious campaign analysis back then needed the patience of a watchmaker and the frustration tolerance of a Vienna taxi driver.

Over those years something fundamental became clear, without anyone saying it out loud: the actual value of our work never lay in the formula itself. Clients paid for the interpretation. For the context no spreadsheet could deliver. But 60% of working time went into the vehicle: into the formula, the preparation, the back-and-forth between export and import. The ratio was grotesque.

Today at SlopeLift we build Agentic Workflows that take over whole pieces of campaign management autonomously: budget shifts between channels, automatic creative rotation based on performance signals, reporting that assembles itself. For someone who spent hours in 2008 patching VLOOKUP formulas, that sometimes feels like science fiction.

Human-on-the-Loop: the decisive difference

There is one client whose ROAS collapses in calendar week 12 every year. Reliably, reproducibly, every time. An agent sees that data point and reacts: cut the budget, adjust the bids, pause the campaign. All logical. All wrong. Because the drop has nothing to do with the campaign. In the target group it is payday week, and buying behaviour normalises by itself afterwards.

That is context knowledge which appears in no dataset. That is the core of Human-on-the-Loop: the AI does the work, the human sets the goals and steps in when context is needed. In-the-Loop means: the human does the work, AI assists. On-the-Loop means: AI does the work, the human steers and corrects.

What will not become a commodity in Austria

In every technology shift (from Smart Bidding to Performance Max to Agentic Workflows) the same question has come up: what is left when execution becomes a commodity? The answer is the same every time.

The technology becomes a commodity. These three things will not. That is the contemporary version of what we meant in 2014 when we called SlopeLift a “workshop”: the efficiency technology delivers is what makes craft genuinely sharp.

The questions that belong on the table now

Whether the title is CAIO or something else: the function has to exist. Someone with budget, mandate and operational understanding who treats AI as a strategic transformation for the whole company.

Three questions for decision-makers in Austria that I consider concrete, not post-AGI speculation: which processes can an agent handle autonomously? Which data do agents need to be able to read, now that text alone is no longer enough since Gemini Embedding 2? Who carries responsibility for AI in the company, with budget and mandate?

The singularity is already happening. As a creeping process, one workflow after another, until we look back and notice that everything has changed. The adoption window is smaller than it was for the jump from offline to online. These are questions for right now.