As artificial intelligence moves from experimental tools into everyday business operations, companies are facing a new challenge: managing an expanding collection of AI models, agents and applications without losing control of their data and workflows. Copenhagen based startup Palette is tackling this problem by developing an operating system that separates a company’s internal knowledge from the AI models it uses, allowing businesses to adopt new technologies without rebuilding their systems every time a better model emerges.
Palette has raised €3 million in pre seed funding to accelerate development of its platform and expand its commercial operations.
The funding round was led by Ugly Duckling Ventures, with participation from Emblem, Acadian Ventures and a group of angel investors.
Founded in 2025, Palette is building infrastructure designed to help organisations integrate AI agents into everyday workflows across departments including sales, operations and product.
Making AI Models Swappable
Palette’s core idea is that companies should not become dependent on a particular AI model or provider.
As new models are released at an increasingly rapid pace, businesses may want to switch between providers based on factors such as performance, cost, speed, privacy or specialised capabilities.
Palette separates these models from the company specific context, workflows and knowledge that make AI useful inside an organisation.
This allows businesses to retain their own operational information while changing the underlying AI systems powering their applications.
The approach is designed to make AI adoption more flexible while reducing the risk of companies becoming locked into a single provider.
A Desktop Environment for AI Agents
One of Palette’s first products is Palette Desktop, an application that allows teams to work with AI agents directly across files and folders.
The platform supports agents including Claude Code and Codex, while giving users the ability to select different underlying models depending on their requirements.
Cloud based models can be selected when teams need greater speed or advanced capabilities, while local models can be used when organisations require greater control over sensitive information.
This flexibility is intended to give employees a practical way to use AI agents within their existing workflows without forcing companies to standardise around a single model.
Palette Desktop is currently available to paying customers, while the company’s broader operating system is being tested with design partners.
Building a Company Owned AI Context Layer
Beyond its desktop application, Palette is developing a broader Palette OS designed to provide organisations with infrastructure for managing AI across their businesses.
The platform includes a company owned context library that stores the information and knowledge AI agents need to perform useful work.
It is also developing shared AI skills that can be reused across different agents, alongside a connector gateway that allows tools to be connected once and made available across multiple AI systems.
By bringing these components together, Palette aims to create a common foundation through which businesses can deploy and manage AI agents across different departments.
Preparing for the Next AI Model
The rapid pace of AI development is creating uncertainty for businesses investing in AI infrastructure.
A model that is considered highly capable today can quickly be replaced by a newer system with better performance or lower costs. Palette believes companies should therefore build their AI infrastructure around durable organisational context rather than individual models.
Its strategy is to make the model interchangeable while keeping the company’s knowledge, workflows and processes consistent.
Expanding the Platform
The new funding will allow Palette to expand its sales operations and grow its team while continuing development of its AI operating system.
The company is targeting organisations that are moving beyond individual AI experiments and looking to deploy AI agents across multiple functions.
As businesses increasingly adopt AI for operational tasks, Palette aims to provide the infrastructure that allows those systems to evolve without forcing companies to repeatedly rebuild their workflows.
With its model independent approach, desktop product and broader operating system under development, Palette is positioning itself as an infrastructure layer for organisations entering the next phase of enterprise AI adoption.