London-based Archestra has raised a $3.3 million Pre-Seed round to build an open-source platform that allows employees to safely connect generative AI tools to internal systems and company data.
The round was led by UK early-stage specialist Concept Ventures, with participation from Zero Prime Ventures, Celero Ventures, RTP Global and Aloniq. The raise, which closed in under two weeks, also attracted a cohort of notable angel investors including Max Hauser (Managing Director & Partner, BCG), Maxim Konovalov (co-founder of Nginx), and Stephen Whitworth (CEO of incident.io).
Founded by serial entrepreneurs Matvey Kukuy and Ildar Iskhakov, Archestra works at the intersection of enterprise AI and data integration.
The company is tackling a major problem facing enterprises experimenting with Large Language Models (LLMs): how to integrate them safely with sensitive company data. AI agents connected to internal tools like Slack, email, CRMs, or HR platforms can unlock huge productivity gains (drafting emails, automating tasks, and retrieving documents autonomously) but they also pose serious risks if left unsupervised.
By layering permissioning, data management, and compliance tools on top of the MCP framework, the platform aims to make safe AI adoption a reality for large-scale organisations.
“An AI agent with full system access could scrape salary data and leak it, or accidentally delete critical internal information,” the company warned.
To mitigate this, Archestra has built what it describes as a security-first orchestrator for the Model Context Protocol (MCP), a framework introduced by AI firm Anthropic in late 2024. MCP enables LLMs to take actions across enterprise systems in a more context-aware and useful way.
“MCP is unlocking a new frontier for AI agents. But, right now, MCP is completely unsuitable for the enterprise,” said CEO Matvey Kukuy.
“We’re building the security-first solution that will change this. Our open-source platform means anyone, whether you’re an engineer or an HR rep, can safely and impactfully integrate AI agents into their workflow, driving efficiency and impact at scale, without increasing risks.”
“Just as APIs became the foundational building blocks for internet infrastructure, MCPs are emerging as the connective tissue for improving the context layer of AI tools within enterprise,” said Ariel Rahamim, Principal at Concept Ventures. “Archestra is building the infrastructure layer this ecosystem needs.”
While major cloud and AI players like Microsoft, Google, and OpenAI are pushing deeper into the enterprise adoption of generative AI, trust and safety remain key barriers to scaling usage. Archestra is betting that its early focus on security, openness, and interoperability will position it as a foundational layer in how companies safely harness autonomous AI agents.
Would you like to write the first comment?
Login to post comments