German AI startup Marble Cognition has raised $3.75M in pre-seed funding to scale its platform that learns how teams work by observation, then deploys AI coworkers to automate manual processes without requiring formal documentation.
The round was co-led by Heartcore Capital and Seedcamp. Among the angel investors were a Datadog board member, Personio’s founder and executives at McKinsey, Mistral AI and Lovable.
Based in Munich, Marble Cognition was founded in late 2025. The company will use the new funding to scale inference, grow its engineering team in Munich and launch in the US.
Felix Rösner, co-founder and CEO, said: "Every company we talk to wants AI to take over real work. But nobody can tell the AI how that work is actually done."
Marble’s platform tracks team workflows and turns them into structured maps AI can interpret. It then builds AI coworkers to carry out each process from start to finish, using the tools teams already have. Rösner said: "Our 'coworkers' watch how the work happens before they take it over."
What we know
- Rösner previously worked at AI sales-agent company 11x.
- Co-founder Christophe Schmit is a former MIT researcher who previously worked at Enam.
- Rösner and Schmit met at Munich's Center for Digital Technology and Management.
Still unclear
- The specific identities of the tech unicorns, private equity-backed firms, and the high-performing US financial services company mentioned in their metrics remain confidential.
- It is not yet known how Marble structures its business model whether it plans to charge per "AI coworker" deployed, via a standard SaaS seat license, or based on volume usage.
Reshma Sohoni, general partner at Seedcamp, said: "Most companies want to automate more, but the hard part has always been mapping out how the work gets done. Marble fills that gap, giving companies vital context for AI transformations that could define their future."
Among Marble’s customers are EGYM, Mutares and Enam, as well as tech unicorns, private equity-backed companies and AI roll-ups. The company says its platform leaves people involved in decisions that require judgement.
Deployment results
At an unnamed US financial-services company owned by private equity, one coworker cut handling time per case by 90% and enabled the team to process 50% more cases in its first month, with no additional hires, Marble said. The coworker learns the workflow by watching the team, then prepares settlement offers for a specialist to approve.
The market gap
About 90 per cent of the core business processes at companies it works with lack formal documentation, Marble said. In the US, automation could theoretically cover approximately 57 per cent of working hours using technology already available, McKinsey estimates.
Max Niederhofer, partner at Heartcore, said: "Marble learns a whole company's processes, anonymises the data, derives SOPs and deploys agentic AI against them."
