Marble raises $3.75M pre-seed to build AI Coworkers that learn by watching work
AI startup Marble Cognition has raised $3.75 million in pre-seed funding, co-led by Heartcore and Seedcamp, with the funds to be used to scale its reasoning capabilities, expand its Munich engineering team and enter the US market. Marble uses a desktop app to observe employees' screen activity, clicks and keystrokes, turning undocumented business processes into structured process maps, and then builds AI Coworkers on that basis that can call on enterprises' existing tools, while retaining a human approval step. Marble says its technology has already been deployed at organizations including EGYM, Mutares and Enam, and that about 90% of the core business processes at the companies it has engaged with have no formal documentation.
Body
AI startup Marble Cognition (Marble) has raised $3.75 million in a pre-seed funding round co-led by Heartcore and Seedcamp, with participation from angel investors including a board member of Datadog, a founder of Personio, and executives from McKinsey, Mistral AI and Lovable.
Founded in late 2025 by Felix Rösner and Christophe Schmit, Marble develops AI-powered technology designed to help companies identify, understand and automate their internal business processes.
As AI technology advances, companies have increasing opportunities to automate operational tasks. However, many business processes remain undocumented, with essential information distributed across emails, spreadsheets and employees' knowledge. This makes it difficult for organisations to identify which activities can be automated and how AI systems should perform them.
According to McKinsey, existing technology could theoretically automate approximately 57 per cent of working hours in the US. Marble reports that, among the companies it works with, around 90 per cent of core business processes are not formally documented, creating challenges for implementing AI-driven automation.
Marble aims to address this problem by analysing how employees carry out their daily tasks and converting these activities into structured process maps that AI systems can interpret.
Based on this information, the platform develops AI agents, known as Marble Coworkers, that can execute business processes using the tools already available within an organisation, while allowing human involvement in tasks requiring judgement or approval.
Every company we talk to wants AI to take over real work. But nobody can tell the AI how that work is actually done. So we solve that first. Our Coworkers watch how the work happens before they take it over. They do it the way your best people do,
said Felix Rösner, co-founder of Marble.
The platform uses a desktop application to observe workflows across business applications, including screen activity, clicks and keystrokes. According to Marble, information is anonymised before being accessed by employees or AI models, and data is analysed in aggregate rather than at an individual level. The application does not monitor private applications, private browser tabs or activity outside working hours, and employees can pause monitoring at any time.
Marble reports that its technology has already been deployed at several organisations, including EGYM, Mutares and Enam, as well as other technology companies and private equity-backed businesses.
The new funding will be used to scale Marble's inference capabilities, expand its engineering team in Munich and support its launch in the US market.