In an opinion column published on SiliconANGLE by Jason Bloomberg, founder and managing director of Intellyx, the author reviews the AGNTCon + MCPCon Europe 2026 conference held in Amsterdam. According to Bloomberg, despite the palpable buzz surrounding agentic artificial intelligence (Agentic AI), attendees were not particularly worried that the technology would lead to catastrophic harm to humanity. While frontier model vendors did not appear on the exhibition floor, the startups exhibiting focused on practical business applications of agentic AI. The author notes that the pace of innovation in this market is outrunning the pace of adoption, yet bona fide paying customer stories already exist alongside new business models developing on a weekly basis in an unquestionably nascent market.
Bloomberg selected seven companies exhibiting at the conference as representative of the balance between customer traction and rapid innovation, dividing the core trends across several areas:
Risk Management and Governance as the Foundation for Agentic Activity
According to Bloomberg, he discounts wild proclamations that AI will grow out of control because of all the effort and innovation invested in controlling it. Two vendors stood out in this area:
-
Traefik Labs SAS: The company leverages its expertise in cloud-native proxy and ingress controller technologies to offer a combined gateway comprising APIs, artificial intelligence, and the Model Context Protocol (MCP). This product essentially consists of three separate gateways sharing a common binary. By monitoring and controlling all the ways AI agents interact with their environment, Traefik delivers a Sovereign Trust Plane—a unified control plane that provides detailed evidence of agent behavior to meet regulatory, auditing, and forensic requirements.
-
Bluerock Security Inc.: The company builds on its pre-AI technology—a military-grade microkernel hypervisor that guaranteed data isolation on bare metal. The company now offers a governed workspace for AI-native builders (both developers and business users working with AI coding agents on their desktops). This workspace acts as an "AI jail" that monitors access to models, manages token usage, and integrates with enterprise identity management systems. In addition, the company offers code scanning capabilities for MCP servers and AI agent skills, extending the field of static application security testing (SAST).
Agent Orchestration: Business Value and Operational Control
When organizations string multiple agents together in an orchestration process, both the potential and the commensurate risk increase. Bloomberg highlights two vendors presenting solutions in this domain:
-
Orkes Inc.: Originating from developments inside Netflix, the company leverages a scalable cloud-native orchestration platform to deliver controlled orchestration in the agentic era. Orkes enables organizations to build deterministic orchestrations, reducing the risk of exceeding operational boundaries, alongside durability and scalability. The platform also includes explainability capabilities, allowing operators to inspect and track interactions between agents.
-
Grape Up Sp. Z o.o.: Best known for its Europe-based cloud and data management platforms, this company offers the Aiboostr platform for AI orchestration and governance. The platform is designed for compliance with the European Artificial Intelligence Act (EU AI Act). This law requires organizations to maintain an inventory of AI capabilities and track the owner, purpose, and users of each agent—tasks managed through Aiboostr's built-in catalog.
MCP-Based Language Model (LLM) Plugins
LLM plugins are distributable workflow packages based on the MCP protocol that can contain reusable skills and applications. These plugins enable models such as ChatGPT or Gemini to return not only text, images, or video, but also plugins that connect to external data and actions:
-
Manufact Inc.: Offers a cloud-based MCP platform that enables developers to build and run components that agents interact with via MCP, including LLM plugins. The platform provides the analytics, testing, and compliance required to deploy MCP-based applications into production.
-
Alpic SAS: The company leverages LLM plugins for business-to-consumer (B2C) customers. These plugins allow functional user interface widgets to be presented directly within the user's chat interface, similar to the portlets used in portal technologies in the past. For example, when a consumer asks ChatGPT for a travel itinerary with specific requirements (such as a hotel room with a particular bed), Alpic enables the model to return a mini-application through which the user can view information and complete the reservation directly from the desktop.
Architectural Discipline for AI-Based Code Development
Bloomberg highlights Reboot Inc. as a notable solution to the challenges of "vibe coding"—a situation where AI coding agents generate all the code for production, creating a challenge in reviewing, understanding, and controlling the generated code. Reboot offers a framework that raises the level of abstraction from the code to the architecture:
- In the Reboot dashboard, developers work directly with business domains (such as "customer" or "invoice") and behaviors, which define what users or agents can do.
- The architectural abstractions in the dashboard constitute a live, working representation of the application code itself.
- The company also provides the backend infrastructure for applications, delivering durable state management, transactionality support, and reactive front ends at cloud-native scale.
The Intellyx Take
Concluding the column, Bloomberg notes that while the market remains in its early evolutionary stages and innovation continues apace, enterprises continue moving forward with agentic AI strategies due to the business value agents promise. According to him, business demand will ultimately drive the market to develop responses to the technological risks of AI agents.