Shaping the Future of Telco Operations with Agentic AI on Amazon Web Services

Global telecom operators are under pressure. Network demand continues to surge while revenues stagnate:

• Revenue Growth: According to PwC, global telecom service revenue grew just 4.3% in 2023 and is projected to grow at only 2.9% CAGR through 2028—below inflation.
• Declining ARPU: Average mobile revenue per user is expected to decline by 1.3% annually.
• Exploding Traffic: Ericsson projects mobile data traffic will double by 2030, driven by video and 5G subscriptions.
• Operational Costs: Global telecom opex decreased by only 0.2% in 2024, leaving little room for efficiency gains.

Meanwhile, AI adoption remains limited. Most operators use generative AI only for chatbots, while advanced use cases such as predictive maintenance or anomaly detection remain immature due to integration and data challenges.

The result: traffic and complexity are accelerating while revenues and margins remain flat. Traditional analytics and incremental automation can’t close this gap.

A New Model: Agentic AI for Telco

The next evolution in telco operations will be driven by agentic AI—a system where multiple intelligent agents collaborate across domains to handle complexity autonomously.

Built on Amazon Web Services prescriptive guidance, this approach enables telcos to:

• Deploy local agents fine-tuned with proprietary data for specific domains (e.g., RAN optimization, transport monitoring).
• Use supervisory agents powered by Amazon Bedrock Agent Core to orchestrate cross-domain operations.
• Ensure compliance and data sovereignty with local deployments on AWS Outposts.
• Enable real-time responsiveness in latency-sensitive functions such as dynamic spectrum allocation or network slicing.

Key Building Blocks of Agentic AI on Amazon Web Services

1. Local Domain Agents

• Fine-Tuned Models: Using Amazon SageMaker and LoRA techniques, telcos can customize open-weight models (8B–30B parameters) for specific operational tasks.
• On-Premises Deployment: Running on AWS Outposts ensures sensitive data remains local, meeting regulatory requirements (GDPR, HIPAA, PIPL, etc.).
• Efficiency: Right-sized models balance performance with computational cost.

2. Supervisory Agents

• Hosted centrally on AWS, supervisory agents coordinate domain agents.
• Powered by Amazon Bedrock Agent Core and optimized with Amazon Nova Pro for reasoning and planning.
• Enable cross-domain orchestration—for example, linking RAN root-cause detection to automated actions across transport, core, and customer service.

3. Inter-Agent Collaboration

• Agent-to-Agent (A2A) Protocol: Enables agents to share insights, validate actions, and collaborate on complex workflows.
• Model Context Protocol (MCP): Standardizes integration with OSS/BSS, IT systems, and network functions, ensuring secure, predictable operations.

This ecosystem shifts telco operations from centralized troubleshooting to distributed, autonomous, and collaborative intelligence.

Benefits for Telco Operators

• Latency & Resilience: Local agents on AWS Outposts provide near real-time optimization and anomaly detection.
• Compliance & Sovereignty: Sensitive data remains within operator premises, meeting strict regulatory standards.
• Customization & Accuracy: Fine-tuned models improve detection, root cause analysis, and resolution speed.
• Operational Efficiency: McKinsey estimates AI-driven automation can reduce opex by 6–9%, with up to 70% of incidents automated.
• CapEx Reduction: Case studies suggest 15–25% savings by eliminating legacy licenses, maintenance, and support.

Next Steps: Path to Autonomous Operations

Telco operators don’t need wholesale transformation overnight. Amazon Web Services recommends a phased adoption strategy:

1. Identify high-value domains such as incident management or predictive maintenance.
2. Pilot distributed agents with measurable KPIs.
3. Scale horizontally across network domains.
4. Layer supervisory orchestration for end-to-end automation.

Operators who embrace this agentic AI model will transform their networks into adaptive, autonomous platforms—ready to deliver next-generation digital services with agility and resilience.

Conclusion

The telco industry is at a turning point. Rising complexity, growing demand, and flat revenues require a new operational paradigm. With AWS, agentic AI offers telco operators the building blocks to create intelligent, autonomous operations—balancing compliance, cost savings, and innovation.

As an official Amazon Web Services partner, Softprom helps telecom operators design and deploy AWS-powered AI solutions that accelerate efficiency, ensure compliance, and prepare networks for the future of digital services.