Munich’s Interloom secures $16.5M in funding
▼ Summary
– Interloom is a Munich-based startup building a “Context Graph,” a continuously updated model of how operational decisions are actually made within an organization, derived from analyzing millions of real cases.
– The company addresses a key problem in enterprise AI: AI agents cannot replicate the unwritten, experiential judgment of long-term employees, which constitutes an estimated 70% of operational knowledge.
– On March 19, Interloom announced a $16.5 million seed funding round led by DN Capital, a significant increase from its initial $3 million seed round the previous year.
– Early customers like Commerzbank use Interloom’s system to analyze support emails and documentation, reportedly reducing the gap between documented procedures and actual work from 50% to 5%.
– Investors back the company based on the thesis that capturing specific organizational context is critical for effective AI agents, representing the next major wave in enterprise automation after robotic process automation.
A major obstacle emerges when deploying enterprise AI agents in large organizations. While technically proficient at following documented procedures, these systems often lack the nuanced, experiential judgment of a seasoned employee. This critical gap exists because a vast reservoir of operational knowledge is never formally written down. Munich-based startup Interloom is tackling this exact problem, recently securing $16.5 million in a seed funding round led by DN Capital, with participation from Bek Ventures and existing investor Air Street Capital.
The company’s solution centers on its proprietary Context Graph. This is not a static database of manuals, but a dynamic, living model of how decisions are truly made within a business. It is built by analyzing millions of real-world interactions, including support tickets, work orders, call transcripts, and emails. By processing this data, Interloom’s system identifies the actual patterns and paths expert employees use to resolve issues, capturing the tacit knowledge that standard documentation misses.
Founder and CEO Fabian Jakobi highlights the scale of the challenge, estimating that roughly 70% of operational decisions lack any formal record. He draws an analogy to Google Maps, explaining that just as navigation apps learn optimal routes from live traffic data, Interloom maps the proven paths experts take to solve problems. This map then guides both AI agents and new employees, creating a continuously updated institutional memory that persists even when individual experts leave the company.
This addresses a pressing demographic shift. With thousands of experienced workers retiring daily, organizations risk losing decades of accumulated expertise just as they attempt to implement AI for complex tasks. Without a way to capture this knowledge, AI systems lack the necessary context to be effective. Interloom’s technology aims to bridge this gap, ensuring that operational knowledge is preserved and leveraged.
Early implementations demonstrate its potential. At Commerzbank, applying the Context Graph to customer support data reportedly narrowed the disparity between documented procedures and actual practice from approximately 50% down to 5%. The startup’s client roster also includes Zurich Insurance, where it won an internal AI competition for an underwriting application, alongside JLL, Volkswagen, and logistics firm Fiege.
The investor group underscores confidence in Interloom’s approach. DN Capital partner Guy Ward Thomas, who led the round, previously backed conversational AI platform Cognigy. His experience reinforced that organization-specific context is paramount for practical AI agent success. Mehmet Atici of Bek Ventures, an early supporter of RPA leader UiPath, views the rise of AI agent adoption as the next major evolution in enterprise automation, following robotic process automation. This funding round, a substantial increase from the company’s initial $3 million seed financing, positions Interloom to scale its mission of making implicit enterprise expertise explicitly usable.
(Source: The Next Web)
