Exclusive: Interloom, a startup capturing ‘tacit knowledge’ to power AI agents, raises $16.5 million in venture funding

$16.5M venture round led by DN Capital (with Bek Ventures and existing investor Air Street Capital) funds Interloom, which previously raised a $3M seed in March 2024; valuation was not disclosed. Interloom ingests operational records to build a continuously updated “context graph” that captures tacit institutional knowledge; deployments at Commerzbank, Volkswagen and Zurich Insurance include a Commerzbank project that narrowed the gap between documented and actual operational knowledge from ~50% to ~5%. The company is pushing a manager-facing “Chief of Staff” agent-management layer; the news is positive for enterprise AI/private markets but likely limited in broader public-market impact.

Analysis

Enterprises trying to automate human-driven workflows will pay a premium for solutions that capture implicit, institution-specific decision rules and put them under version control; that creates a data-asset moat more durable than any single LLM fine-tune. Expect a typical customer journey from pilot to production to take 6–24 months, driven by integration, validation by domain experts, and legal sign-off — not by model accuracy alone.

Cloud and platform economics are a key second-order effect: continuous ingestion and query of historical operational records materially upsells high-margin storage, vector search, and inference hours to hyperscalers and SaaS vendors, lifting TAM-per-customer by an order of magnitude versus one-off automation scripts. Conversely, consultancies that monetize repetitive process knowledge are exposed to revenue compression unless they transition to engineering and retained-data-services models.

Main downside paths are operational: noisy or conflicting historical records, PII/regulatory constraints on reuse, and a single high-profile hallucination or compliance incident that forces conservative rollbacks; any of these can push payback periods beyond 24 months and re-open the inertia advantage. Key short-term catalysts to watch are enterprise pilot conversions, regulatory guidance on internal data reuse, and marquee acquisitions of dataset-rich startups, each capable of re-pricing public integrators and cloud vendors within 3–12 months.

Key Decisions for Investors

• Long PATH (UiPath): buy a 9–12 month call spread to express asymmetric upside from platform-style adoption (allocate ~1% NAV). Rationale: RPA incumbents can capture orchestration value and datasets; target +40% upside if enterprise pilots scale, max loss = premium paid (~100% of premium).
• Pair trade — Long NOW (ServiceNow) / Short ACN (Accenture): equal-dollar exposure over 12–24 months. Rationale: platform owners that embed agent governance win recurring revenue while consultancies lose low-value process work. Target outcomes: NOW +30% / ACN -20%; set stops at -12% on either leg.
• Long GOOGL (or GOOG): buy 12-month calls or overweight cloud exposure (1–2% NAV). Rationale: sustained ingestion and vector query demand should drive incremental cloud spend even if models are on-prem; payoff >2x if adoption accelerates, downside limited to option premium or small equity drawdown if hyperscalers bundle solutions.
• Tactical event play: on any credible acquisition rumor of a context-graph startup, tactically buy PATH or NOW on the assumption these integrators become acquirers/partners; hold 3–9 months for a takeover-driven rerate, sizing small (0.5–1% NAV) to limit idiosyncratic event risk.