Topics AI (Artificial Intelligence), Autonomous networks
Autonomous Broadband delivering seamless integration between wholesale networks and retail providers, enabling zero-touch operations and superior end-to-end customer experience
Project companies
Autonomous Broadband Orchestration
The Problem
In open-access broadband markets, last-mile fibre is delivered by a wholesale provider (NBN) while retail services are managed by a separate CSP (TPG). This structural separation was designed for regulatory fairness, not operational efficiency. The result: fragmented provisioning, misaligned data models, and heavy manual coordination across organisational boundaries.
Customers experience the consequences directly. Service activation takes days, not hours. Faults trigger multi-team escalation journeys spanning Care, NOC, and Network teams. WiFi issues are misidentified as network problems, driving unnecessary support calls. Accountability disappears at the wholesale-retail boundary, and the customer bears the cost of every handoff, delay, and miscommunication.
The Solution
AutoNeX introduces an Agentic AI-driven cross-operator orchestration layer that makes the wholesale-retail boundary operationally invisible. Service requests are expressed as high-level business intents and dynamically executed by AI agents that reason across both domains in real time.
The system operates on a continuous closed-loop cycle: Monitor, Detect, Decide, Act, Validate. When remediation does not resolve an issue, agents adapt their approach. This is not faster troubleshooting. This is autonomous operational intelligence.
Customer Impact
AutoNeX eliminates customer friction at every touchpoint:
* Activation: 30-50% faster service provisioning, moving from days toward near real-time. Order fallout reduced 30-50%, eliminating the "your order failed" experience.
* Issue Resolution: Agentic AI performs fault demarcation in seconds during live customer interactions, replacing hours of manual triage and multi-team handoffs. First-contact resolution improves dramatically.
* Proactive Prevention: The system detects degradation before customers notice. Technicians are dispatched before impact. WiFi issues self-heal autonomously. The best customer experience is the one that never becomes a support call.
Operational Gains
* 70-80% of cross-operator workflows execute without manual intervention
* 40-60% reduction in manual operational effort (OPEX)
* 30% improvement in Mean Time to Resolve (MTTR)
* 20-40% fewer SLA breaches
* Evidence of closed-loop, intent-driven behaviour: the system continuously translates business intents into cross-domain actions, validates outcomes, and self-corrects without human escalation
Scalability
The architecture is designed to scale and improve simultaneously:
* AWS serverless infrastructure (Bedrock, AgentCore, Lambda, Step Functions, EventBridge) scales elastically from thousands to millions of transactions with zero capacity ceilings
* Splunk Cloud ingests telemetry from growing customer bases, with predictive models improving as data volume increases
* AI agents learn continuously from operational data. Churn predictions sharpen, fault demarcation accelerates, WiFi remediation becomes more precise over time
* Cost-per-customer decreases with volume as fixed platform costs amortise and automation depth compounds
* The unified cross-operator data model built on TM Forum Open APIs (TMF641, TMF638, TMF642, TMF921) and ODA ensures any new market deployment is a configuration exercise, not a rebuild
The wholesale-retail structural separation exists in 20+ countries globally. AutoNeX provides a replicable, standards-based blueprint deployable in any open-access market worldwide.
Method and Confidence
Three concrete use cases demonstrate the approach works in practice:
1. Proactive Churn Reduction - Predictive monitoring of customer experience indicators (REXI scores below 30) triggers autonomous technician dispatch before customers are impacted
2. Agentic AI Dark NOC Triage - Real-time fault demarcation across NBN Access Network, NTD, and TPG WiFi CPE during live customer interactions, with minimal human intervention
3. Self-Healing WiFi Experience Management - Continuous monitoring of WiFi health indicators with autonomous remediation (dynamic channel optimisation, interference mitigation) before issues escalate
Each use case delivers consistent, repeatable outcomes at scale: improved speed, reduced manual steps, and measurable customer satisfaction uplift across the key journeys that define broadband experience.
The Outcome
AutoNeX proves that structurally separated operators can function as one autonomous system. It delivers better outcomes for customers, lower costs for operators, and a scalable path toward industry-wide autonomy.
Autonomy is the engine. Customer experience is the outcome.
Resources
Resources
AutoNeX: Autonomous Broadband Orchestration - Team Members Videos
Kuldip Tripathi - TCS TM Forum Assets
Chris Kumar - AWS The Technology Platform
Tomás Iglesias Melendez - Datadope The Data Intelligence
Sam Voukenas - Splunk Cloud The Data Foundation
Demos
Use Case 1 - Proactive Churn Reduction
Use Case 3 - Proactive WIFI Experience Management
Use Case 2 - Reactive Triage Improvement
AutoNeX - End to End Demo
Infographic
Project summary infographic
AutoNeX: Autonomous Broadband Orchestration
Presentations
TPG AutoNex Catalyst Main Presentation
TPG AutoNeX Catalyst ARENA Presentation
TPG AutoNeX Catalyst ReXI Presentation
Solution Design
Solution Diagram
AutoNeX in Action - Solution Design
AutoNeX: Solution Design Explanation
TM Forum Assets
TM Forum Assets
Contact team
Email the members of the Catalyst team to request more details.