AI has fundamentally changed the nature of knowledge work, continuously challenging organizations to rethink their operations. At trivago, we’re leveraging AI to maximize the potential of our 600 person team. Our ambition is to empower our 600+ core talents to operate with the impact of 6,000.
- trivago isn’t starting from zero: AI has run in production for over a decade, and more than 60% of the workforce already thinks in systems, builds with code, and works in feedback loops.
- Roughly 43% of trivago’s collective working hours today go to execution work AI can progressively absorb, but the real leverage lies in each of us becoming dramatically better at the work only humans can do, sharper decisions, faster building, deeper judgment, gains that compound into something efficiency alone can’t produce
- Every talent is expected to map their own repetitive work, build new skills, and treat AI fluency as continuous craft-leveling, while leaders are expected to model the shift
- trivago provides its talents with the tools, frameworks, and time to build AI fluency, real training in self-leadership, resilience, and mental flexibility, and, above all, the space to explore and create value for users, the business, and each other
- Internal survey data backs the momentum: 93% of talents now use AI daily, up from a fifth having never used it as recently as 2023, and 84% feel confident adapting to it.
- As work evolves, trivago has identified four archetypes, Navigator, Builder, Orchestrator, and Connector, which form an interdependent flywheel for how human value gets created as AI absorbs execution.
The last two and a half years have brought an unprecedented acceleration to the working world, changing the nature of knowledge work in real time. For a company like ours, this is an exciting opportunity. We are investing in ourselves, actively reshaping how we work, and building a more resilient and effective trivago for whatever comes next. You have probably seen the headlines. Company X is betting big on AI, shedding thousands of roles and calling the savings on salaries an “efficiency gain.” This has become the default story of our time: technology in, people out.
Ours is a different story.
We’re building an AI-native organization, one that uses the best of technology to empower its roughly 600 core talents to operate with the impact of 6,000. Not by our people being replaced with AI, but by the work that slows them down being removed, by each of them being freed to become dramatically better at the work only humans can do.
We’ve been tracking AI adoption internally for four years, and the numbers reflect real evolution. In 2023, 22% of our talents told us they had never used AI. Today, that number is zero. 93%* now use AI daily or multiple times a day. But adoption alone doesn’t equal value. What is more telling is that 80% of our talents say they’re genuinely excited about what AI means for their role, and it’s already translating into something concrete: talents report saving an average of 55 minutes a day, 26 working days a year. What interests us most is how that time is spent. Several respondents said they take on more, or more complex, work because AI enables them to do things they couldn’t do before.
Across the company, teams are redesigning their workflows around AI, rebuilding them from scratch, and sharing what they learn with each other. While leadership is steering the direction, the energy driving this change comes from our talents. This transformation is happening with us, not to us. We’re shaping it, and like any real transformation, that means learning as we go without having all the answers yet
trivago is not new to this
For two decades, we’ve built our reputation in travel search by staying ahead of the technology curve, not catching up to it. AI has been part of that story since long before this current shift.
The ranking that decides which hotel a traveler sees first. The algorithms behind our bidding and marketing spend. The personalization running quietly under every search. All of it has run on production AI for more than a decade. We shipped one of the first AI-powered semantic search experiences in the category, across multiple languages, in 2023. It means travelers can search however they naturally think, and still get an intuitive search experience on trivago. We’ve also had a lot of fun innovating outside of product. In 2023, we became the first to launch an AI-powered TV campaign, and we’ve kept pushing the envelope of what is possible with AI in marketing and creative production since: production time has dropped significantly, and our creative teams now own more of the process end to end.
The developments in agentic AI over the past year have us genuinely excited, not only for what they unlock in our product, but for what they unlock in our organization. The maturity of agentic AI has broadened our ambitions, and we are convinced that a different scale of leverage and sharper competitive edge is within our reach.
We are throwing out the old model: humans execute, technology assists. We’re building our organization around a new one: AI-driven systems execute, humans direct. The first gives you speed. The second gives you learning loops, systems that improve through their own output, feeding what they learn back into how they run. That’s the real difference between a company that uses AI and a company that runs on it.
Our speedboat advantage
Most conversations about AI assume that scale is the advantage. In this transition, the opposite can just as easily be true. We share an industry with giants and the platforms built on top of them. Their size is also their constraint. Enormous headcount, legacy systems, and layers of process make it slow to turn something that large in a genuinely new direction, even when the intent is there.
trivago is a speedboat by comparison: sizable enough to matter, nimble enough to actually move. We’re curious and adaptable by habit, with a genuine track record of reinvention. That combination lets us experiment and change direction quickly, without the weight of a large headcount to carry or protect. And because we’re built predominantly of builders rather than people who manage process, our central question right now isn’t how to shrink. It’s how fast we can turn what we’re learning into momentum, and that’s a far more energizing question to spend three years answering.
At our size, we’re big enough to build real infrastructure and small enough to keep a culture built on curiosity and trust, where the instinct is to learn fast and execute faster. That happens to be much of the behavior AI adoption rewards. We believe the real opportunity isn’t doing today’s work a little faster. It’s asking what our systems would look like if we designed them from a blank page, with agents owning execution end to end: building things AI-first by design, rather than bolting AI onto workflows we built for humans. As a company of builders, that’s exactly the kind of problem we’re wired to solve.
Efficiency is the floor. Craft is the ceiling.
There’s a version of this argument that treats any job AI touches as a job on its way out. We think that gets it backwards. Every time a system takes over the part of a job that was never the interesting part to begin with, the person who used to do it isn’t left with less. They’re left facing a harder, more valuable problem than the one they were solving before.
It would be reasonable to assume a company chasing 10x impact is quietly planning to get a lot bigger. We’re doing close to the opposite. The number on our org chart is the one thing we are not trying to change. What we are trying to change is what that number can do.
Here is where most discussions of AI transformation go wrong, and where we want to be precise about the math. Across trivago, roughly 43% of our collective working hours currently go to execution work that AI can progressively take on. Automating a meaningful share of that work, bringing it down toward 15%, is genuinely valuable. It’s roughly the equivalent of adding 150 people to the organization without a single new hire. But if that were the whole story, we would simply be a slightly more efficient version of the company we already are. That is the floor, not the ceiling.
The genuine leap comes from somewhere else. As AI takes on execution work, trivago talents gain more than time back. They get sharper at the tasks a human is genuinely best placed to handle: better decision-makers, faster and more ambitious builders, people who can hold far more complexity in view and who have room to deepen the relationships that actually move the business forward. When each person becomes meaningfully more effective at their craft, those gains compound into something larger than any efficiency saving could produce on its own. Efficiency multiplies capacity. Craft multiplies outcome.
“Automating the mundane work is not where the leverage is. The leverage comes from becoming better builders, better orchestrators, better connectors, from how we level up our human craft. Efficiency gives us more capacity, but the multiplication comes from our craft. That is how 600 people reach the impact of 6,000.” – Johannes Thomas, CEO, trivago
Our talents feel ready for it. 84% say they feel confident adapting to AI, and that number matters as much as any usage stat, because talent, not tools, is what actually drives this transformation forward: the instinct to spot an opportunity, the curiosity to try it, and the follow-through to turn it into value for themselves, for travelers, and for our partners.
What changes most is where human judgment sits. Where people once worked in the loop, executing each task by hand, one at a time, they increasingly work on the loop: directing systems, reviewing their output, catching what needs catching, and improving how the whole thing runs. The human contribution doesn’t get smaller in this picture. It moves to where it’s worth the most, and it grows there.
Craft is only one half of this, though. The other half is what happens to the workflow itself. Multiplying a person’s skill inside a process built for a different era only gets you so far. The bigger unlock comes from taking that process apart and rebuilding it with AI at the center from the start, instead of adding a smarter assistant to the same sequence of human steps. That is slower, harder work, and it is also where the real 10x lives: not just people getting better at old workflows, but old workflows getting replaced by ones that couldn’t have existed without AI in the first place.
Usage dashboards and token spend make for an easy scoreboard, but they don’t tell you where value is actually being created. Every company will soon be able to claim high adoption; the models themselves are becoming a commodity faster than anyone predicted. What won’t be common is the harder work underneath: who actually gets to make a call, how something one person learns reaches the other 599, where a judgment call hands off to a system and back again. Watch that, not the usage stats, and you’ll know within a year or two which companies actually changed and which ones just added a new tool to an old way of working.
Where Human Effort Goes Next
As AI absorbs more of the execution layer, the real question is where human effort goes instead. We look at it through four archetypes: a practical lens on the kind of value creation a machine can’t replicate on its own, and one we can actually build action from.
- A Navigator sets the direction and frames the problem worth solving.
- A Builder constructs the system or pipeline that solves it.
- An Orchestrator governs what’s running and keeps the quality bar high.
- A Connector holds the trust and alignment that keeps the other three actually working together instead of in silos.
This isn’t a new org chart or a job-title overhaul. It’s a deliberately simple, holistic way of talking about where a person’s effort is going: a shared language rather than a classification system.
Every role at trivago carries some mix of all four archetypes, plus whatever slice of pure execution AI hasn’t yet absorbed. No two people, and no two teams, carry the same mix. That’s exactly why this framework exists: not to sort people into boxes, but to give everyone a common way to talk about how their work is shifting and where it’s headed. The four archetypes also don’t work alone. They form something closer to an engine, each one leaning on the others to actually produce value. Knowing your own mix is where the real thinking starts: what to double down on, what to build toward, and where your biggest room to grow actually sits.
Builder is trivago’s dominant archetype today. Over 40% of our people work in Technology, a structural advantage most companies our size don’t carry into this transition. We expect the Orchestrator role to evolve and become even more important over the next few years: as AI systems multiply across every part of the business, the need for someone to govern them grows just as fast, and it’s already the archetype with the largest projected growth ahead of it, which makes it both our biggest shift and our biggest re-skilling opportunity. Our teams are already organized around owning problems end to end rather than handing work down a chain, exactly the structure this next phase rewards. Much of the foundation is already there. The work now is connecting it, speeding it up, and extending it everywhere.
Job titles rarely keep pace with the work itself, and that gap is only going to widen. A role can look identical on paper while the tasks inside it quietly move somewhere else: hours once spent on manual reporting now going into shaping the prompts that generate it and reviewing what an agent produces before it goes out. None of that shows up on an org chart, and none of it waits for a title update first.
That’s the real ask for anyone doing knowledge work right now: look past what your job title says today, and pay attention to what your role is actually becoming. Waiting for the title to catch up before you adapt is a losing bet, and it’s not something any employer today can hand you fully formed. The requirement is self-leadership: taking ownership of moving toward wherever the leverage sits, treating learning new skills as consistent habit rather than an annual training slot, staying teachable in all aspects, and staying focused on the problem in front of you, not the one that mattered two quarters ago.
We built a direct answer to the obvious next question, “where do I fit in this?”: an AI Role Evolution Agent, developed in-house. It walks a trivago talent through a short, guided conversation about what their day-to-day work actually looks like, then produces a personal report: which of the four archetypes above their role leans toward today, how much of their current work AI can realistically absorb, what their role could plausibly look like three years out, and concrete next steps to get there. It gives talents a shared language for growth conversations as much as a personal roadmap.
Our culture is the multiplier
The technology itself is, in the end, the easy part. Before long, every company in our industry will have access to roughly the same models. What can’t be bought, and where our real advantage lives, is the culture and mindset to use them well.
Over the past three and a half years, we’ve learned and upskilled together, because affinity with a new tool was never going to be equal across the company, and we recognized that early. What’s carried us through is trivago’s culture of fanatic learning: a growth mindset that meets change with curiosity instead of resistance, real agency over your own growth rather than wait to be told how, and a habit of sharing what you learn so the whole company gets smarter together. trivago runs formal training too, but that’s not what’s actually driving this. What’s driving it is our talents meeting this transformation with curiosity and determination, choosing to recreate their own work with AI at the center: to create value, to innovate, and to build the capacity for the harder tasks that are inherently human. The pace of that learning and building has been extraordinary.
That shows up in habits you can actually watch happening. AI Days that individual disciplines have adapted to their own work. A cross-functional group of AI Ambassadors, drawn from nearly every part of the business, who help colleagues turn curiosity into working tools rather than waiting for a central “AI department” to hand down solutions. An Agentic Marketplace, the one-stop place to discover agentic components already built and in use across trivago, so nobody has to start from zero. And, from this year, something more structural: both team and individual developmental OKRs now carry AI learning and upskilling on equal footing with every other performance expectation. Learning isn’t squeezed in beside the job anymore. It is part of the job, as seriously and holistically as anything else on a talent’s plate.
None of this asks trivago’s people to become someone they’re not. It asks them to lean further into what has always been rewarded here: learn fast, execute faster. That entrepreneurial instinct is really a form of self-leadership: taking ownership of a problem and finding a solution, rather than waiting for a directive.
That expectation doesn’t stop at individual contributors. It’s reshaping what we ask of leaders too. AI is increasingly doing the coordinating work middle management used to do: routing information, flagging what needs attention, closing gaps between teams. That doesn’t shrink leadership’s role, it raises the bar. The leaders who matter here aren’t the ones managing through layers. They’re the ones close enough to the work to act on it directly, and increasingly, leading fleets of agents alongside their teams.
“Companies and people who want to win with AI have to become highly liquid. That means moving talent quickly to where the leverage actually is. It means learning continuously, not in occasional training bursts. It means pointing people at the problems that matter right now, not the ones that mattered last year, or the ones that will matter next year. It asks a lot of people and leaders too: a willingness to let go of work you were good at, to learn in public, to be measured on outcomes rather than on the tasks you used to own. The companies that make that safe, even energizing, are the ones that will pull ahead.” – Johannes Thomas, CEO, trivago
None of that holds without leaders finding their own footing first. In a period with this much genuine uncertainty, you cannot give a team direction you haven’t found for yourself. Self-leadership isn’t a personal development extra here. It’s the precondition for leading anyone else at all.
We ask something concrete of everyone at trivago in return: invest one hour every day pushing AI to its limits in real workflows. Stay curious. Keep learning. This is essential, not optional.
Building the flywheel
Individual habits don’t scale a company on their own. Someone learning a new tool in their spare time is a good start. But it isn’t a transformation. What actually compounds is a structure that turns individual learning into organizational capability, over and over, without depending on any one person to carry it.
We think about that structure as four layers, each translating the transformation into something concrete for a different scope: the company, a craft, a team, and a single person.
-
The Company Playbook sets the direction: the four archetypes, the ambition to take 600 people to the impact of 6,000. It’s owned by the leadership team.
-
A Discipline Playbook takes that direction into a specific craft, such as software engineering, data science, product, or UX. It’s written once by a thought leader(s) in that discipline and used as the shared starting point by every team with people in that craft.
-
A Team Playbook brings it down to one team’s actual workflows, commitments, and OKRs: a concrete, multi-year path for the work that team does, authored by the team lead with their team, not handed to them.
-
And at the individual layer, the AI Role Evolution Agent provides a clear personal picture of how AI is already reshaping a talents work. Looks at where automation is today and where your role and archetype blend is heading in the next years. This turns the company-wide direction into a personal view of where someone’s own craft is heading.
None of this is being handed down from the top and left for teams to execute. It’s being built together, by talents and leaders alike, with all of it pointed at the same outcome: a more resilient, more competitive trivago for the people who use it, invest in it, and work here.
Each piece feeds the next. A skill one team builds gets documented, added to a growing shared skills library, picked up by another team through the Playbook layer above it, and improved along the way. It’s the same self-improving loop we’re chasing in our technical systems, just built at the level of the organization instead of a single workflow. Learning turns into tooling. Tooling turns into shared capability. Shared capability turns into outcomes. And outcomes fund the next round of learning.
Moving fast without losing the plot also takes a shared set of rules. Five principles guide how we work through this transformation: how we move, measure, scale, guard, and share.
We’re just getting started
trivago has changed shape several times over two decades, always the same way: by learning faster than the world around us was changing. That instinct is exactly what this moment rewards.
This won’t arrive as a single before-and-after moment. It will move through the company at different speeds in different places. At any given time, trivago will look like a hybrid: some teams several steps ahead, others just getting started, all moving in the same direction. That unevenness isn’t a gap to close overnight. It’s what a company actually looks like mid-transformation.
We don’t have a crystal ball, and we’re not pretending otherwise. What we have is a structured approach to learning and adapting rather than a fixed plan to execute. The archetypes, playbooks, and AI Role Evolution Agent are all built so that individual learning turns into organizational capability. We’re energized by what’s ahead, not because we’ve figured it out, but because we haven’t. Working that out in the open with our people is worth sharing. We plan to continue telling this story honestly as it unfolds, sharing what worked and what didn’t.
We’re confident that this will make trivago more competitive and innovative because our people, not just our tools, are getting sharper. That means a workplace where our talents are genuinely challenged, where they grow, and where they can be their best selves. That’s harder to quantify than a statistic, but it’s just as real. That’s what being the fastest learning and executing company in our space actually looks like from the inside, not the speed alone, but people who are better for it.
Follow us on LinkedIn to stay updated on our milestones and learnings.
About trivago N.V.
trivago N.V. (NASDAQ: TRVG) is a leading global hotel search and price comparison platform, and one of the most recognized travel brands in the world. When travelers search for a hotel, we want trivago to be the obvious choice. We help them find the best place to stay and deliver the best deal to book, saving them time and money – so every traveler feels smart and confident about their booking. Powered by AI, we personalize and simplify hotel search for millions of travelers, connecting them with more than 7.0 million hotels and other accommodations across more than 190 countries.