24/7 Customer Service

Industrial AI, IOT and Automation

Qualified Expert Professionals

ORION – Orchestrated Reinforcement-learning Intelligent Organisational Network of LLM Agents. IRIS has been backed by the UK Sovereign AI programme to develop advanced Large Language Model technologies for secure, sovereign, and scalable enterprise and industrial AI applications.

ORION is an Innovate UK-funded sovereign AI project developing a secure multi-agent Large Language Model system for enterprise and industrial decision support. The platform is designed to help organisations analyse information, generate reports, interpret structured data, support decision-making, detect anomalies, and automate knowledge-intensive workflows, while keeping data under stronger organisational control.

Developed as a TRL-4 proof-of-concept, ORION brings together several core technologies, including an intelligent coordinator/orchestrator agent, multiple specialist LLM agents, retrieval-augmented generation, safety and governance mechanisms, intelligent synthetic data generation, reinforcement-learning-based improvement workflows, and human-in-the-loop validation.

The architecture supports local and on-premise deployment, making it suitable for organisations that require stronger data sovereignty, traceability, auditability, and control over sensitive information.

ORION includes multiple specialist agents for tasks such as administration, scheduling, finance, SQL/data queries, chart interpretation, reporting, and system evaluation. These agents are coordinated by a central orchestrator, which decomposes user requests, selects the appropriate agents and tools, combines outputs, and provides structured responses.

The project also includes a Lab Mode, which allows new agents to be configured, tested, validated, and improved before deployment. This supports future extensibility, safer system evolution, and continuous improvement of the platform.

During the project, IRIS successfully delivered the integrated ORION proof-of-concept and validated the system across key performance areas including accuracy, safety, auditability, anomaly detection, data interpretation, agent coordination, and user trust.

ORION has the potential to improve enterprise productivity, reduce reliance on external AI providers, support secure AI adoption, and create a scalable foundation for future sovereign AI solutions across multiple sectors.