From AI models to AI systems: agents, multiagents, orchestration (I)

21 October 2026 | 09.05h to 10.00h

Artificial intelligence is evolving from models capable of analyzing data to predict, prescribe, or generate new content into more complex AI systems made up of agents that interact with one another, with tools, and with digital environments. In this session, four examples will showcase how AI agents are being applied across different sectors, including the legal and healthcare fields.

  • Xesca Alabart Lopez, CEO, Easyspecs

  • Albert Iruela Sánchez, CEO and Founder, Omnios

  • Oscar Cardona Garrido, Head of AI Solutions, Sopra Steria

  • Toni Alonso Solís, Head of the Artificial Intelligence Area, Hospital General de Granollers

Speakers:

XESCA ALABART LOPEZ

CEO, Easyspecs

Xesca Alabart is founder of easyspecs.ai, an agent-driven platform for automated specs documentation validation and generation. As an Agentic Coding Coach and Solutions Architect, she helps teams and organizations transform software development through AI-native architectures and Requirements-Driven Development patterns. With 15+ years of CTO-level experience in fintech, insurtech, and media (AMINA Bank, Bolttech, Dorna Sports/MotoGP), she is a specialist in agentic infrastructure, OpenCode agents, and AI Factory methodology.

Building 100x More Productive Development Teams with AI-Native Architectures

What multiplies a development team’s productivity exponentially? It’s not just tools—it’s architecture, methodology, and cognition. In this session, we’ll explore how Compound Engineering, as the methodological skeleton, structures AI-native architectures, Coding Harness platforms (OpenCode, Claude Code, etc.), and Requirements-Driven Development. The key innovation: Dark Factory—instead of coding a feature, we code a factory that generates the feature automatically. We’ll cover real-world cases, patterns that work, and how to apply AI Factory methodology to generate specs, validate code, and iterate in shorter cycles. Result: teams that ship 100x faster.

ALBERT IRUELA SÁNCHEZ

CEO & founder, Omnios

Albert Iruela Sánchez trained as a telecommunications engineer and has previous experience in technology consulting. He is the founder and CEO of Omnios, a company specialised in the adoption of artificial intelligence in corporate environments and public administrations. As CEO of Omnios, he drives results-oriented AI projects, from strategic definition to the development and productisation of solutions. Among his current responsibilities, he supports organisations in the implementation of specialised solutions for natural language processing across text, voice and video, while helping to minimise risks throughout the process.

Lexia: from an AI model to an orchestrated system for legal document management

This presentation will introduce Lexia, the AI agent developed by Omnios with Bonpreu to transform legal document management. Starting from a real challenge, querying critical, scattered and unstructured legal documentation, it will explain how the solution evolved from a document chatbot into an orchestrated AI system, combining specialised agents, API integration, daily information updates, source traceability and deadline automation. The talk will show how organisations can move from AI models to robust, useful systems deployed in production.

OSCAR CARDONA GARRIDO

Head of AI Solutions, Sopra Steria

Oscar Cardona is Head of AI Solutions at Sopra Steria Spain, where he leads the design, development, and deployment of artificial intelligence solutions that drive digital transformation across key sectors of the economy. With over 15 years of experience in software engineering and technological innovation, he has progressed from product development to leading generative AI initiatives, combining a strategic vision with strong technical expertise.

Agent Orchestration and Living Knowledge with the Nexyro Platform

How do we move from isolated AI models to systems that are integrated into the daily workflows of development teams? In this session, we will explore Nexyro, Sopra Steria’s generative AI platform that transforms project knowledge into a living, actionable asset.
We will demonstrate how it creates a living knowledge base from sources such as code, documentation, video, and audio, and how, through agent orchestration, this knowledge is continuously updated and leveraged across the entire software development lifecycle—from automated documentation generation to team support and onboarding.

TONI ALONSO SOLÍS

Artificial Intelligence Manager, Hospital General de Granollers

Toni Alonso és responsable de l’Àrea d’Intel·ligència Artificial de l’Hospital General de Granollers. Enginyer informàtic, compta amb més de 15 anys d’experiència en enginyeria del programari, transformació digital i lideratge de projectes tecnològics. Actualment lidera iniciatives d’IA generativa aplicada a la salut, centrades en sistemes d’IA basats en RAG, agents i arquitectures orquestrades, amb l’objectiu de millorar l’accés a la informació clínica, donar suport a la presa de decisions i optimitzar els processos assistencials.

RHCIA: An Orchestrated AI System for Automated Clinical Summarization and Intelligent Access to Catalonia’s Shared Electronic Health Record

RHCIA is an orchestrated AI system that transforms Catalonia’s Shared Electronic Health Record (HC3) into an intelligent clinical information access tool for healthcare professionals. Triggered by clinical events such as emergency admissions, the system automatically retrieves relevant patient data, generates structured clinical summaries, and enables conversational querying through a RAG-based approach with source traceability. This session will demonstrate how orchestration, generative AI, and information retrieval can be combined to move beyond standalone models and build complete AI systems that support real-world clinical decision-making.