Munich-based startup Interloom secured $16.5 million in funding.

      The Munich-based startup is developing what it refers to as a ‘context graph,’ an ever-evolving map that illustrates how operational decisions are made within a business, based on millions of actual cases rather than potentially non-existent documentation.

      There is a specific friction point in every implementation of enterprise AI, and those who have attempted to deploy an AI agent in a large organization typically encounter it early on. While the agent may possess technical capabilities, such as reading documentation, adhering to instructions, and performing tasks, it cannot replicate the intuition of an employee who has been in the role for fifteen years and understands, through practical experience, why the established procedures may falter on certain days, such as Tuesdays in logistics. This nuanced knowledge has never been documented, as there was no incentive to do so until now.

      This is the challenge that Interloom aims to address. The Munich startup, which operates in Munich, Berlin, and London, announced on March 19 that it secured $16.5 million in a seed round led by DN Capital, with additional participation from Bek Ventures and existing supporter Air Street Capital.

      This funding round marks a substantial increase from the company's initial $3 million seed round, which Air Street Capital led in March 2024 when the company first came out of stealth mode.

      Interloom’s primary offering is known as the Context Graph: an ever-changing model that captures how operational decisions are actually made in a specific organization by analyzing millions of authentic cases, including support emails, service tickets, call transcripts, and work orders, thereby revealing the patterns through which experienced workers resolve issues.

      Founder and CEO Fabian Jakobi discusses the issue in terms of tacit knowledge, a term introduced by British-Hungarian philosopher Michael Polanyi, who observed that most expertise cannot be fully expressed by the expert themselves. Jakobi estimates that approximately 70% of operational decisions are not formally recorded.

      He likens the process to Google Maps: just as the application learns optimal routes based on live traffic, Interloom constructs a representation of the pathways that operational experts utilize to solve problems, subsequently employing that representation to assist AI agents and new employees when confronted with similar challenges. The system is perpetually updated, allowing every resolved case to contribute to the institutional memory instead of fading away when the individual who resolved it departs or retires.

      This retirement risk is integral to the startup's pitch. The press release references a statistic indicating that 10,000 baby boomers are exiting the US workforce daily, a demographic trend well documented by Pew Research.

      The assertion is that organizations are grappling with a compounding issue: the loss of decades' worth of institutional knowledge is occurring precisely as AI is anticipated to automate complex operational tasks. If that knowledge is not first captured, the AI has no valuable foundation to rely on.

      Interloom’s initial clientele includes Zurich Insurance, JLL, and logistics company Fiege, along with Commerzbank and Volkswagen, both of which were independently confirmed by Fortune in its exclusive report on the funding. At Commerzbank, Interloom examined millions of customer support emails in relation to internal documentation and reportedly narrowed the disparity between what was documented and how work was actually conducted from approximately 50% down to 5%.

      At Zurich Insurance, the company triumphed in an internal AI competition against what Jakobi described to Fortune as 2,000 rival AI-native startups for a specific underwriting use case.

      The lineup of investors also lends support to its potential. Guy Ward Thomas, the DN Capital partner overseeing the investment, was previously the first institutional investor in Cognigy, a German enterprise conversational AI platform, which DN Capital supported since its Series A in 2019 and was acquired by NICE in August 2025 for $955 million—marked at that time as the largest AI exit in Europe.

      Ward Thomas has emphasized that the key takeaway from that investment was the importance of organization-specific context in ensuring the successful operation of AI agents in practice. Mehmet Atici, who leads Bek Ventures' participation, was an early supporter of UiPath, the robotic process automation pioneer that went public in New York in 2021. His assertion is that the current trend of adopting AI agents signifies the next major shift in enterprise automation following RPA.