In collaboration with IBM Consulting, trivago builds and scales “trivago Copilot” that improves productivity, knowledge access, and decision making
The launch of generative AI marked a turning point in how work is created, documented, and executed across industries. Knowledge workers suddenly gained new ways to draft ideas, search for information, and accelerate decision making. At the same time, organizations faced growing uncertainty around access to models, data governance, and how to safely enable these tools at scale. On top of it, trivago needed to address strict European data protection and digital sovereignty requirements. The use of public AI tools raised concerns about sensitive company information being processed or transferred outside the EU, limiting the ability of teams to freely experiment with generative AI. This created an urgent need for a secure, enterprise-grade environment where employees could safely interact with AI while ensuring data residency, governance, and compliance with EU regulations. The risk was not only adopting the wrong technology but also moving too slowly while the nature of work was already changing.
Within this broader context, trivago recognized that its existing challenges would become more visible without a new approach. The market was moving fast, with different models excelling at different tasks. The overhead of managing separate subscriptions, logins, and mental contexts was eroding the very productivity gains AI was supposed to deliver. At the same time, consolidating around a single provider was not a safe alternative: pricing changes, model deprecations, or service interruptions could disrupt workflows with little notice, creating concentration risk that no organization can ignore. These challenges created an opportunity for trivago to establish a secure and unified interface for interacting with enterprise data using AI, simplifying access to distributed information while enabling the safe, compliant, and scalable use of generative AI within controlled environments.
Creating high quality documentation, project plans, and business cases required significant manual effort, slowing execution in an otherwise fast moving, entrepreneurial environment.
trivago needed a way to respond to the generative AI shift while improving productivity, preserving knowledge, and empowering employees at all levels to work faster and make better decisions.
To turn generative AI into a reliable enterprise capability, trivago partnered with IBM to design and evolve its internal AI copilot, called “trivago Copilot”. Through IBM Consulting’s Digital Product Design and Engineering, IBM and trivago collaborated to build a flexible, open foundation capable of supporting multiple large language model providers while maintaining governance, security, and regional data control. The trivago Copilot experience aligns with modern enterprise copilots, leveraging a cloud foundation to deliver secure, scalable AI-driven productivity.
The teams established a consistent and automated way to provision and manage the underlying data and integrations, enabling faster iteration and reliable scaling across the organization. This foundation allowed trivago to expand the copilot beyond content generation and evolve it into an enterprise search experience that connects emails, Slack messages, and documentation systems, making knowledge easier to retrieve regardless of where it resides. This capability was further enhanced through retrieval-augmented generation (RAG), ensuring that AI responses are grounded in trivago’s internal data while keeping all information securely within the controlled environment. This approach reinforced digital sovereignty by preventing sensitive data from being exposed to external systems.
To extend beyond knowledge retrieval, trivago implemented secure, AI-driven automation of daily workflows. Employees can now execute tasks such as sending emails through Outlook, summarizing Slack conversations, and creating or updating Jira tickets through simple prompts. This allows teams to move from passive information consumption to active task execution using AI. The platform also leverages AI agents to execute code, generate and download files, create images, and interact with multiple file formats, further expanding the scope of what employees can accomplish through a single interface.
To move from platform to practice, trivago launched the Copilot alongside a company-wide AI Learning Day: 25 sessions built around real use cases, attended by approximately 70% of the company.
Rather than limiting AI to a small group of specialists, trivago focused on broad enablement across the workforce. A dedicated AI Ambassador Group helped define objectives, share best practices, and guide employees through hands-on experimentation, while leadership actively reinforced a culture of curiosity, learning, and responsible use of AI.
As Carolina Muradas, Head of AI, Strategy and Operations Lead at trivago explained, “The risk of not adopting AI is much higher than the risk of learning by doing. By working with IBM, we were able to move from concept to production faster than we would have alone. Now our teams start projects from a higher perspective.”
By working with IBM, trivago ensured that AI adoption moved beyond experimentation, embedding AI into everyday ways of working rather than treating it as a standalone initiative. All capabilities were designed with EU data residency in mind, ensuring that enterprise data, generated outputs, and user interactions remain securely within the region.
Within the first years of using trivago Copilot, employees reported saving an average of eight working days per year on the first year, in 2023. Today, more than 600 employees actively use the platform, with adoption reaching approximately 90% of the eligible workforce, demonstrating strong integration into daily operations. In the third year, in 2025, that figure more than doubled to over 16.6 days per employee, reflecting broad adoption and deeper integration of the team with trivago Copilot.
Beyond time savings, trivago strengthened how knowledge is captured, structured, and reused. Enterprise search reduced rework by making past decisions, documentation, and context easier to access, while teams improved the overall quality of documentation, planning, and decision making. Employees can now start initiatives from a more informed baseline, with clearer visibility into existing insights. Beyond knowledge access, employees can now automate routine workflows and execute actions directly through the platform, reducing manual effort and freeing time for higher-value work.
As a result, trivago shortened the time from idea to execution, enabling faster experimentation and iteration without sacrificing quality. Importantly, trivago achieved this transformation while maintaining full control over its data, ensuring compliance with European regulations and establishing a trusted, secure environment for enterprise AI adoption. The outcome is not only higher productivity, but a more consistent and resilient way of working that allows the company to adapt as generative AI continues to evolve. Akshay Sureka, Associate Partner at IBM Consulting, summarises: “The real impact comes from embedding AI into everyday workflows. By connecting enterprise knowledge and enabling action through a single interface, trivago teams can move faster with greater confidence.”
trivago is a global technology company that helps travelers compare hotel prices and find the best accommodation options. The company operates with a strong culture of entrepreneurship, transparency, and continuous innovation.
Legal
© Copyright IBM Corporation. June, 2026.
IBM, the IBM logo, and IBM Product Engineering Consulting are trademarks of IBM Corp., registered in many jurisdictions worldwide. Examples presented as illustrative only. Actual results will vary based on client configurations and conditions and, therefore, generally expected results cannot be provided.