Home
...
GSMA Agentic AI Summit at MWC Barcelona 2026
Agentic AI Summit Overview
The Agentic AI Summit cut through the hype to deliver a sharp, four-hour briefing on how Agentic AI—systems capable of complex, goal-oriented execution – was set to revolutionise Telecoms Networks and End-User Services.
As 5G-A infrastructure begins to mature, this summit addressed the critical questions for MNOs and the ecosystem: How does the industry move from automated processes to truly autonomous network operations? How can network capabilities securely and profitably exposed to external AI agents to unlock new consumer and enterprise revenue streams?
Over four sessions, leaders explored the four pillars of agentic adoption:
- Network Autonomy: Deploying agents for self-healing and operational efficiency.
- Agentic Value Drivers: How agents can create top line value for the industry, not just cost-savings.
- Industry-wide Challenges: Getting to grips with the challenges facing Agentic AI within telecoms networks.
- The Trust Layer: Defining the security and governance standards for a fully agent-driven ecosystem.
Over 800 attendees gained actionable insights into monetisation strategies, infrastructure evolution, and ecosystem development needed to lead in the adoption of AI for networks to understand how Agentic AI could lay the groundwork for an intelligent and globally connected mobile infrastructure.
The Agentic Imperative
To kick-off the summit, the audience sat down with Dr Junlan Feng from one of the world’s largest and most advanced mobile operators (China Mobile – CMCC). This session examined why Agentic AI was not just an incremental improvement, but an urgent strategic imperative for the telecommunications industry. This visionary conversation framed the day’s discussions by connecting the promise of autonomy to the core challenges of monetisation, efficiency, and future relevance.
Speaker: Dr. Junlan Feng, China Mobile
ServiceNow: Turning AI Investment into Growth Velocity
When it came to AI, telcos want to unlock the next billion in revenue, not just better operations. Being efficient kept them alive, but being autonomous was what allowed them to grow; the session argued that growth was impossible when networks were disconnected from the rest of the service experience. Real autonomy only happened when AI could see across the front, middle, and back office, act on what it learned, and move fast enough to capture growth before it disappeared. Attendees learned how operators were putting AI to work with ServiceNow to move at the speed growth demanded, and why zero-touch networks were table stakes to the growth conversation.
Speaker: Rohit Batra, ServiceNow
Agentic AI: Network Autonomy
The dream of a “zero-touch” network is becoming a reality through Agentic AI. This session dived into the practical application of AI agents within the network core to create a self-diagnosing, self-healing, and continuously optimising operational environment. Leaders from foundational AI and workflow automation platforms shared how operators were moving from pilot projects to full-scale deployment, unlocking significant OpEx savings and unprecedented reliability.
Chair: Guy Lupo, TM Forum
Speakers: Lilac Ilan (Nvidia), Alex Jinsung Choi (Softbank), Olivier Simon (Orange), Rohit Batra (ServiceNow)
WIT Software: When AI takes the call
This session explored how telecom operators could evolve beyond traditional connectivity and leverage the new wave of agentic AI. It examined how AI agents could become first-class subscribers, owning phone numbers, operating across multiple channels, and delivering intelligent communication services. It highlighted how this transition allowed operators to unlock new revenue opportunities and position themselves for the emerging Agentic AI population.
Speaker: Luis Silva, WIT Software
Agents in Action
This session explored how telecom operators could monetise Agentic AI by transitioning from connectivity providers to high-value service engines. Participants examined the deployment of autonomous agents to deliver personalised consumer engagement and intelligent enterprise solutions, such as virtual front desks and sector-specific smart operations. The discussion covered practical use cases for autonomous voice and digital interactions that redefined the customer experience, unlocked sustainable revenue streams, and maximised lifetime value.
Chair: Louis Powell (GSMA)
Speakers: Martin Kon (Cohere), Luis Silva (WIT Software), Jesús Martin Tello (NTT Data)
Open Telco AI Assets -The Importance of Open Telco models
This discussion focused on the importance of Open Telco models and how key players such as AT&T and AMD are forging their way, alongside the GSMA, in curating open telco models, data and benchmarks for the accelerating the adoption of AI within the Telco Community.
Speakers: Mark Austin (AT&T) and Alex Sinclair (GSMA)
AWS: Bridging the AI Chasm
Generative AI and agentic AI opened unprecedented opportunities for revenue growth. Industries like healthcare and finance were already capitalising on new services and personalised cross-selling at scale. For the telecom industry, this represented a transformative moment. While leveraging AI for cost reduction delivered benefits, the session argued that the real breakthrough lay in AI-enabled innovation. This session explored how telcos could bridge the AI chasm to capture top-line growth, examining how agentic AI systems autonomously identified opportunities and enabled innovative services on cloud infrastructure.
Speaker: Ishwar Parulkar, Amazon Web Services
NANDA: Resolution, Discovery, and Governance for the Agentic Web Research at MIT
As AI agents became autonomous and long-lived across devices and the cloud, the core challenge shifted from agent design to trust and governance at internet scale. NANDA (Networked AI Agents in Decentralised Architecture) addressed this by introducing a systems-level framework for the Agentic Web. It focused on agent resolvability, certification, and discovery, enabling agents to be verifiable and safely composable across network boundaries. This session explored how NANDA worked on technologies envisioned for the 6G era.
Speaker: Ramesh Raskar, Massachusetts Institute of Technology (MIT)
A Fundamental Change, Thriving in a Post-Human Network
With the prediction that autonomous agents would soon generate more network traffic than humans, this session addressed the immense challenges and opportunities involved. Participants debated the new rules for an agent-first world, redefining the “user” and architecting for machine-to-machine communication at a global scale.
Speakers: Usman Javaid (Orange), Häkan Dursun (GSMA)
Totogi: The five myths of Agentic AI
Danielle Rios challenged the five myths currently driving telco AI strategy—from the “data lake delusion” to the belief that agents alone could work across siloed systems. She revealed that agentic AI only succeeded when grounded in industry context rather than just raw data, warning against expensive demos that lacked P&L impact.
Speakers: Danielle Rios (Totogi / TelcoDR), Sean Kinney (RCR Technology)
Securing the Trust Layer for an Autonomous World
This closing panel addressed the fact that autonomy without trust was impossible. The discussion tackled the toughest challenges: managing identity for billions of non-human agents, liability for agent mistakes, and defence against AI-driven threats.
Chair: Alex Leadbeater, GSMA
Speakers: Kal De (Nokia), Daniel Cuthbert (Banco Santander), Chema Alonso (Cloudflare), David Palmer (Pairpoint by Vodafone)
Palo Alto Networks : Secure AI by Design
The shift to Agentic AI required a fundamental rethink of enterprise security. Attendees joined Palo Alto Networks to explore “Secure AI by Design,” a framework for adopting autonomous agents. The session uncovered unique attack vectors—including Memory Poisoning and Privilege Compromise—and offered a concrete strategy for securing the entire AI lifecycle.
Speaker: Simoné Gammeri, Palo Alto Networks
