Ugly Duckling Ventures led the Palette pre-seed round, with participation from Emblem, Acadian Ventures, and a group of angel investors. Founded in 2025 by Brian Kyed, Lars Ettrup, Steffen D. Sommer, and Christian Lomholt, the Copenhagen startup is stepping into a critical market gap. As AI adoption moves beyond engineering departments and into sales, operations, and product teams, organizations are struggling to manage a fragmented landscape of disconnected AI tools.
A Model-Agnostic AI Operating System
The startup’s core proposition is separating underlying AI models from a company’s internal context, workflows, and knowledge. This architecture allows businesses to switch between different AI providers and models seamlessly, retaining their own organizational data and processes throughout the transition. Teams can optimize their tool selection based on specific factors like speed, cost, or required capabilities.
“There will always be a newer model. The responsible way to adopt AI is to make the model swappable and make your company context durable. That’s what we’re building with Palette,” said co-founder Brian Kyed. This philosophy directly addresses the rapid, unpredictable pace of AI development. By treating the AI model as an interchangeable component rather than a foundational dependency, companies avoid being locked into a single vendor’s ecosystem.
Solving Enterprise AI Fragmentation
The rapid proliferation of AI tools has created operational headaches for modern enterprises. Sales teams might rely on one model for drafting communications, while operations teams use another for data analysis, and product teams experiment with coding assistants. This fragmentation leads to duplicated efforts, inconsistent data handling, and security vulnerabilities.
Palette’s approach centralizes the control layer. By maintaining a unified context library, the company ensures that all AI agents operate from the same foundational company knowledge, regardless of the underlying model powering them. This creates a durable environment for AI-native teams to operate securely and efficiently.
From Desktop App to Enterprise Platform
Palette is currently generating revenue through its Palette Desktop application. This tool enables teams to work with AI agents such as Claude Code and Codex directly on their local files and folders. The application gives users the flexibility to choose between cloud-based models for faster processing and advanced capabilities, or local models when handling sensitive information that demands greater control and privacy.
Alongside the desktop tool, the company is building a broader platform. This system features a company-owned context library, shared AI skills, and a connector gateway. The gateway is particularly focused on efficiency: it allows external tools to be connected once and then utilized across multiple AI agents. This reduces technical debt and accelerates the deployment of swappable AI models across the organization. While the Desktop app serves paying customers today, the broader OS is actively being tested by design partners.
What Happens Next
Palette plans to use the €3 million injection to expand its sales operations and grow its team. The capital will also fund the continued development and rollout of its broader operating system. As the startup scales, its success will depend on convincing enterprises that a separate orchestration layer is necessary for their AI workflows.
Looking forward, the market for infrastructure supporting AI-native teams is poised for significant growth. Companies are realizing that their internal context and workflows are their true competitive advantage, not the specific AI model they license. If Palette can successfully deliver a durable, model-agnostic architecture, it will position itself as a critical utility in the enterprise software stack. Businesses will be able to adopt new, more powerful AI models the moment they are released, without disrupting their established operational processes.
— David Kim, technology desk, AXO News