Skip to main content

Trusted Digital Services: When AI Agents Rely on Network APIs

Published on November 12, 2026

4 min

Chief Technical Officer Working at Metal Components Production Factory. Young Middle Eastern Specialist Walking in the Facility, Using Tablet Computer to Monitor the Manufacturing Process
  • Networks are becoming programmable: innovation in networks is no longer limited to areas such as throughput and latency; the direct use of certain telecom network capabilities offered by APIs now enables the emergence of new applications and services.
  • New use cases are emerging: identity verification, fraud prevention, location services, and on-demand quality of service are paving the way for safer and more personalized experiences.
  • Standardization is crucial: the GSMA Open Gateway and CAMARA initiatives enable developers to access the same capabilities across multiple networks and foster the emergence of a global market.
  • AI could accelerate this transformation: autonomous agents will require reliable information and real-time action capabilities, making the network an essential and secure component of the digital economy.

For decades, mobile networks have evolved to offer ever-greater throughput, coverage, and responsiveness. Today, thanks to network APIs (Application Programming Interfaces), applications can now directly leverage the network’s capabilities in real time. From a simple connectivity infrastructure, the network is becoming a platform for intelligent services that can be activated in real time.

Making Network Capabilities Available on Demand

This transformation is driven by several major technological advances. The virtualization of mobile networks, the maturity of cloud-native architectures, and the new capabilities of 5G enhance network programmability by enabling the management and configuration of network capabilities via software, while exposing them to third parties in a secure and controlled manner.

At Orange, we are already developing and implementing practical applications of this programmability. Since May 2026, our Number Verification API—integrated notably into Google Firebase’s Phone Number Verification service—has made it possible to verify directly with the network that a mobile number corresponds to the line in use on a device, without requiring an SMS code. The goal is to simplify authentication while strengthening the fight against fraud.

Our ambition is to provide third parties—including agents—with secure, real-time access to certain network capabilities for use cases such as identity verification, enhanced and personalized customer experiences, and Industry 4.0.

To this end, in Spain, we launched a 4G/5G on-demand quality-of-service API in July 2026, paving the way for differentiated connectivity experiences tailored to specific needs. The challenge is to transition from a network with predefined performance levels to one where certain capabilities can be requested based on usage.

Capable of performing tasks and making decisions, AI agents will need reliable information about their environment and ways to interact with it. Network APIs are capable of fulfilling this dual role.

Common Standards to Scale Up

For APIs to be widely adopted, they must be able to operate across different networks without requiring developers to implement specific integrations. This is the role of the GSMA Open Gateway and CAMARA initiatives, to which Orange contributes in order to build a global ecosystem of network APIs based on common standards.

To accelerate this movement, we launched Orange LiveNet in 2024, a business unit dedicated to the development and commercialization of these new capabilities. Together with Ericsson, we also helped create Aduna, a joint venture bringing together several major global telecommunications groups to accelerate the development of network APIs on an international scale.

Agent-based AI: The New Driving Force Behind Network APIs

The rise of agent-based AI is opening up new possibilities. Capable of performing tasks and making decisions, AI agents will need reliable information about their environment and ways to interact with it.

Network APIs are capable of fulfilling this dual role. In a factory, for example, an automated system could detect an incident, deploy inspection drones, and request a high-bandwidth connection to transmit real-time images.

At Orange, we are already exploring these possibilities in remote monitoring, remote operations, and differentiated connectivity, with a view toward more autonomous, efficient, and secure industrial processes.

By 2030, the success of network APIs will be measured less by the number of available interfaces than by their adoption by businesses, developers, and platforms. Our ambition is to make the network a trusted foundation for digital interactions, whether human-driven or AI-driven.

Author