Backing Archestra: Building the Guardrails for AI adoption in Enterprise

We’re excited to announce our investment in Archestra, the open-source platform enabling enterprises to securely connect AI agents to their data without the risk of them “going rogue.” The company has raised $3.3M pre-seed in an oversubscribed round we’re proud to have led, alongside Zero Prime Ventures, Celero Ventures, RTP Global, and participation from exceptional angels, including leaders from BCG, incident.io, and ElevenLabs.

Why we invested

The emergence of the MCPs (Model Context Protocol) is unlocking an entirely new class of AI applications. Initially introduced by Anthropic, MCP is to AI what APIs were to the early internet… the connective tissue that allows LLMs to securely integrate with an organisation’s tools, systems, and data sources. This unlocks huge productivity potential, but also introduces serious risks: a poorly managed AI agent could scrape confidential data or delete vital information in seconds.

Archestra is building the infrastructure layer for this ecosystem. An MCP orchestrator that provides guardrails, dynamic permissions, and context management. The platform makes it possible for any employee, technical or not, to safely integrate AI agents into workflows while meeting enterprise-grade security and compliance standards

We were drawn to three core factors:

1. A founding team with proven depth
   CEO Matvey Kukuy and CTO Ildar Iskhakov have been building in open-source and infrastructure for over a decade. They previously co-founded Amixr (acquired by Grafana) and, in Matvey’s case, Keep (acquired by Elastic). They’ve scaled products to millions of users, built vibrant OSS communities, and navigated complex enterprise integrations.
   
   
2. A mission-critical problem, lived first-hand
   Both founders have experienced the challenge of scaling secure, integrated systems in high-stakes environments. They understand that MCP adoption at most enterprises isn’t just about connectivity. It’s about governance, control, and trust.
   
   
3. Executional intensity
   In just three weeks of operating, they assembled a world-class founding engineering team, closed their pre-seed in under two weeks, and committed to an open-core strategy designed to build community adoption and enterprise conversion - check out their GitHub for more updates on https://github.com/archestra-ai/archestra
   
   

The road ahead

The integration-platform-as-a-service (iPaaS) market is forecast to exceed $17B by 2028, and MCP is set to be a foundational part of it. Just as APIs became the backbone of internet infrastructure, we believe MCP will underpin how enterprises deploy and govern AI context at scale. Archestra’s approach of open-source transparency, combined with enterprise-grade guardrails, positions them to define this category.

As our Principal Ariel Rahamim put it:

“From the moment we met the team, it was clear there is a unique blend of operational excellence and a genuine hunger to build something truly category-defining. Having lived and breathed open-source products and GTMs for the last 10 years, it was obvious to us we had to work with Matvey and Ildar in defining the journey over the next 10 years.”

We’re thrilled to welcome Matvey, Ildar, and the Archestra team to the Concept Ventures family, and we can’t wait to support them in building the guardrails for AI’s next frontier.

‍