According to Telefónica’s Alex Harmand, agentic AI changes how network APIs generate revenue by shifting pricing away from individual API calls
Telefónica and Nokia have recently collaborated to test how AI software agents could support the use of network APIs, as part of the GSMA Open Gateway initiative to standardize access to telecom network capabilities. Initial testing has focused on a fraud prevention use case in a lab environment. Telefónica is using Nokia’s Network Exposure Platform to expose network APIs and related functions — including SIM swap and device swap — while Nokia’s Network as Code platform aggregates those capabilities for application developers.
The collaboration reflects a broader industry shift toward agentic AI, where autonomous software can discover, combine, and monetize network capabilities without direct developer orchestration.
RCR Wireless News caught up with Telefónica’s Core and Platforms Senior Manager Alex Harmand to discuss the top use cases — beyond fraud detection — that will drive the most meaningful revenue from agent-enabled network APIs.
Here’s what he said:
Identity and Trust Orchestration. An AI agent can dynamically integrate number verification, SIM status, device context, and location data to generate a real-time contextual identity confidence score. “This supports onboarding, account recovery, and step-up authentication with clear monetization through reduced fraud losses and lower friction,” said Harmand.
Context-aware risk scoring for payments and commerce. In this use case, the agent evaluates transaction context in real time and selects the most relevant trust signals dynamically, rather than relying on static fraud rules, which Harmand said enables “adaptive authentication flows and risk-based pricing models” that directly improve approval rates and reduce chargebacks for payment providers and e-commerce platforms.
Predictive Intelligence-as-a-Service. Prediction-driven agents move beyond simple validation toward proactive optimization. At MWC, Telefónica and Mavenir will showcase a No-Show Prediction agent that forecasts attendance likelihood and triggers targeted interventions to reduce missed appointments. According to Harmand, this use case has clear monetization potential across healthcare, retail, field services, and logistics, where no-shows directly impact revenue.
Network quality and SLA optimization for enterprises. For B2B customers, an AI agent could interpret intent — such as ensuring optimal connectivity for a specific application — and dynamically evaluate factors like latency, congestion, slice availability, and policy constraints. While still early-stage, it represents a promising future use case.
“This is particularly relevant for cloud gaming, media streaming, IoT fleets, and mission-critical enterprise services,” said Harmand, adding that this revenue model is tied to premium connectivity tiers, SLA-backed services, and dynamic quality-on-demand offerings.
Selling outcomes
Big picture, explained Harmand, agentic AI changes how network APIs generate revenue by shifting pricing away from individual API calls toward outcome-based consumption. Instead of billing per endpoint request, operators monetize each agent execution — meaning customers pay for a completed objective, such as a fraud assessment, identity verification, or no-show prediction.
Behind the scenes, an agent may orchestrate multiple APIs and network capabilities, but commercially it is sold as a single, productized action. This simplifies pricing while allowing operators to capture greater value from higher-level automation.
The model also enables operators to package capabilities into bundled services rather than exposing standalone APIs. Offers such as Identity & Trust or Commerce Risk Scoring combine multiple signals and functions into one solution, creating clearer commercial propositions for enterprises.
In practice, agentic APIs shift monetization from selling technical access to selling outcomes — improving customer predictability while increasing average revenue per user by pricing the solution itself rather than each underlying microservice.
“The customer still gets a predictable commercial model, and we capture more value because one agent run may orchestrate multiple underlying capabilities,” said Harmand.
