AI agent environments and multi-agent orchestration

Cargoffer offers consortium partners its expertise in designing, building and operating AI agent environments for transport and operations use cases. We build production-grade tooling that lets LLM-based agents interact with real-world logistics systems, and we have first-hand experience operating multi-agent workflows in production for internal and customer-facing use cases.

We contribute expertise in:

• MCP (Model Context Protocol) server design and operation, with production reference implementations across three logistics verticals — route optimisation (transcend-mcp-server), electronic consignment notes (ecmr-mcp) and freight marketplace (bolsa_de_carga-mcp) — all open-sourced under github.com/cargoffer.

• Multi-agent orchestration patterns: planning agents that decompose logistics tasks, tool-using agents that call our APIs, verification agents that check outputs, and human-in-the-loop checkpoints where the cost of a wrong answer is high.

• Tool-use design for transport-specific LLM agents: how to describe a routing tool so an agent picks the right vehicle profile, how to constrain tool calls so an agent does not invent POIs, and how to surface uncertainty back to the user.

• Interactive agent UIs: a Skybridge MCP App that renders API results in a structured, type-safe UI inside the chat, which is useful for stakeholder-facing demonstrators and for showing the agent's reasoning in a reviewable way.

• LLM-provider-agnostic deployments, including local LLM and self-hosted model options for projects with EU data-residency requirements.

• Operational practices for agent environments: tool versioning, evaluation harnesses, regression tests, prompt libraries and observability for agent traces — all of which matter when an agent is integrated into a real freight workflow.

We are open to contributing this expertise to Horizon Europe 2026-2027 consortia that include AI-agent or copilot components in their work programme. We are particularly relevant to Cluster 4 (Digital & Industry, including the international cooperation in AI call) and to Cluster 5 Mobility projects that need an AI-agent layer to make transport data and optimisation services accessible to non-technical users.

Relevant Horizon Europe calls: HORIZON-CL4-2027-04-DIGITAL-EMERGING-06; HORIZON-CL5-2027-03-D5-03; HORIZON-CL5-2027-03-D5-04; HORIZON-CL5-2027-03-D5-05.

Fields of expertise: Artificial Intelligence / Machine Learning; Information and Communication Technologies; Software / SaaS; Transport and Mobility; Open Source