Qontext raises $2.7M to build a reusable ‘context layer’ for enterprise AI

Qontext, a Berlin startup founded in 2025, has raised a $2.7 million pre‑seed round led by HV Capital to build what it calls an independent context layer for AI.

The round included participation from Zero Prime Ventures and a group of founders and operators across automation and enterprise software, with angels such as Jan Oberhauser (n8n), Emil Eifrem (neo4j), Bastian Nominacher (Celonis), Philipp Heltewig (Cognigy) and Fabian Veit (make.com).

AI model capability is advancing fast, but many firms still struggle to get consistent outcomes when applying AI across business functions. Qontext argues the real blocker isn’t models — it’s fragmented, out‑of‑date context spread across CRMs, docs, product databases and bespoke systems.

'Putting a great model into an organization without context is like expecting a world‑class hire to deliver on day one without any onboarding — the capabilities are there, but the results won’t be,' Lorenz Hieber, Co‑founder & CEO of Qontext, told TechInBerlin. 'With Qontext, companies can roll out new AI tools and agents that are fully context‑aware from day one.'

Qontext’s pitch is a simple architectural shift: treat context as a standalone, reusable layer that is independent from models and apps. Instead of rebuilding context for each new agent or workflow, teams connect to a single, governed source of truth that supplies the right information when an AI process runs.

The company says customers — from fast‑growing startups to larger enterprises — are already using Qontext in production across marketing, sales and support to deliver more consistent, automatable outcomes. According to the startup, centralizing context can more than double the number of processes that can be reliably automated with AI.

'Context fragmentation is one of the toughest infrastructure problems in AI today, and Qontext is solving it at scale,' Jan Oberhauser, Founder & CEO of n8n and an angel investor in the round, told TechInBerlin. Ann‑Christin Stiehl, Investor at HV Capital, added: 'What convinced us is that Qontext is not another AI feature, but a foundational layer every serious AI stack will need.'"

The technical challenge is significant. Qontext will have to keep millions of data points synchronized, resolve conflicting versions of truth, and enforce complex access controls for both humans and autonomous agents.

'We’re dealing with millions of data points, constantly changing information, and complex access controls across humans and agents,' Nikita Kowalski, Co‑founder & CTO of Qontext, told TechInBerlin. 'But solving this is also the biggest lever for making AI work at scale.'

The new funding will go to expanding the platform and hiring engineering and product talent in Berlin to accelerate development of reusable context infrastructure and integrations that let AI processes run on trusted, continuously updated context.

Qontext’s approach highlights a broader trend: after a burst of experimentation with models and LLM features, organisations are now building the surrounding infrastructure — retrieval, memory, and context — needed to scale AI beyond pilots. If Qontext can deliver reliable, governed context at scale, it could remove a common bottleneck and make enterprise AI deployments materially more productive.