What a day at AI SUMMIT BARCELONA — 900+ people, 50+ talks, 4 stages at Yellow Park (Glovo HQ). The energy was next level. Highlights from the stage: Executive Perspectives on AI (w/ TravelPerk, Holaluz, Factorial): real talk on challenges, opportunities, and the future of tech leadership.
Alex
CTO AIDroid Hub
Our LAM & LAM-Agents insights by AI Droid Hub:
From chat to action
2026 will favor AI that does, not just advises. LAM-agents execute across apps and interfaces—end to end.
1
ROI over model metrics
Wins are measured in EBITDA—–30–50% manual ops, –30–40% support load, +7–12% conversion (typical targets from our pilots).
2
Governance by design
Human-in-the-loop, audit trails, scoped permissions, and safe “action sandboxes” are must-haves for enterprise rollouts.
3
Start small, scale fast
Pick 1–2 high-impact processes, prove value in 30–45 days, then expand to a portfolio of agents.
4
Integration without re-platforming
GUI-level actions + APIs mean you can automate today’s stack without long IT projects.
5
Large Action Models (LAM): The Next Generation of AI after ChatGPT or AI is reinvented
Large Action Models (LAMs) are emerging as the next big leap in artificial intelligence – one that moves beyond ChatGPT-style chatbots into AI agents that take action. Unlike today’s large language models (LLMs) that can only understand and generate text, LAMs can actually perform tasks in digital environments at a user’s request. In other words, they turn AI from a passive assistant into an active doer. Experts are calling LAMs the most exciting development in AI since ChatGPT and even more because they unlock a new world of automation and efficiency.