Marble Cognition, a Munich-based startup developing AI coworkers for business operations, has raised €3.35 million ($3.75 million) in a pre-seed funding round. The company says its technology first learns how work is performed inside organisations before helping automate it.
The round was co-led by Heartcore and Seedcamp, with participation from angel investors including a Datadog board member, the founder of Personio, and executives from McKinsey, Mistral AI and Lovable. Marble plans to use the funding to scale inference, expand its Munich engineering team and support its launch in the United States.
Founded in late 2025 by Felix Rösner and Christophe Schmit, who met through Munich’s CDTM programme, Marble says it maps how employees carry out processes and converts that information into structured systems that artificial intelligence can interpret. It then deploys Marble Coworkers, which are designed to handle manual white-collar tasks across the software tools companies already use.
The startup argues that the main obstacle to business automation is often not the availability of AI, but the lack of documented information about how work is actually done. According to Marble, around 90% of core processes in the companies it works with exist only in inboxes, spreadsheets and employees’ knowledge.
Its system observes team workflows and turns each process into a structured map. Marble Coworkers can then execute those processes from end to end, while leaving people involved where human judgement is required. The company distinguishes this approach from traditional automation, which typically adapts a standard product to existing workflows; Marble says it instead builds a system around the process itself.
Marble also described privacy measures for its technology. The company says information is anonymised before it is viewed by people or AI models, including when the system runs in a customer’s own cloud. Its desktop application records screen activity, clicks and keystrokes in workplace applications, while excluding private apps, private browser tabs and activity outside working hours. Employees can pause the recording, and the data is analysed in aggregate rather than at an individual level.
The company’s customers include EGYM, Mutares and Enam, as well as other technology unicorns, private-equity-backed organisations and AI roll-ups supported by major funds.
Marble reported that one AI coworker deployed at a US financial services company owned by a private equity firm reduced case-handling time by 90%. During its first month, the team handled 50% more cases without adding staff. The coworker prepares settlement offers for approval by a specialist, having learned the workflow by observing the team rather than relying on existing documentation.
Sources: EU-Startups.