By Miha Ušeničnik, Associate Director at DFG CONSULTING
The telecommunications landscape is undergoing a fundamental metamorphosis. For decades, network operations have relied on a blend of manual intervention, tribal knowledge, and semi-automated workflows. Today, the industry is pivoting toward a paradigm defined by autonomy: networks that do not merely carry traffic but actively analyze conditions, make split-second decisions, and execute autonomous recovery.
However, as we push toward the vision of "Self-Healing" and "Self-Configuring" infrastructure, a silent crisis is brewing in the back office. While the algorithms governing these networks are becoming increasingly sophisticated, the foundational data—the "ground truth" of the physical network—remains fragmented, outdated, and fundamentally untrustworthy.
The Current State of Autonomous Ambition
The promise of autonomous networks is clear: greater operational efficiency, reduced service provisioning latency, and near-instantaneous fault recovery. According to a landmark study by the IBM Institute for Business Value and the TM Forum, titled Navigating autonomous networks, the industry is acutely aware of the necessity of this transition. Approximately 73% of network executives report that their organizations have established formal, phased roadmaps to move toward autonomous operations.
Yet, there is a stark disconnect between ambition and execution. Currently, only 6% of Communications Service Providers (CSPs) report operating at Level 4 autonomy—a level where the network is capable of autonomous, AI-driven decision-making with minimal human oversight. While 22% of executives expect to reach this milestone within the next three years, the industry faces a significant hurdle. Bridging this gap is not merely a matter of upgrading AI software or deploying more robust telemetry; it is a challenge of data integrity.
The Chronology of Disconnect: A Legacy Burden
To understand why autonomous networks struggle, one must examine the evolution of telecom data over the last thirty years.
The Era of Siloed Evolution (1990s–2010s)
In the early days of network expansion, documentation was often handled via local, fragmented systems. As operators expanded through mergers, acquisitions, and rapid technology cycles (from Copper to DSL to Fiber), network data was siloed. Information was scattered across disparate formats: CAD files for physical layouts, Visio diagrams for logic, PDFs for site surveys, and Excel spreadsheets for inventory management.
The Rise of "Shadow Documentation" (2010s–2020s)
As central inventory systems struggled to keep pace with rapid field deployments, engineers developed a secondary, unofficial system of record. These "shadow documents"—often stored on personal drives or local servers—became the "true" source of information for field technicians. This created a dangerous bifurcation: the official system of record became a repository for stale data, while the actual, functional network existed only in the heads of engineers or in scattered, unmanaged files.
The Automation Bottleneck (Present Day)
We have now reached a stage where automated systems are being plugged into these legacy environments. An AI agent, designed to optimize traffic routing or isolate a fault, relies on the data provided by the central inventory system. If that data is inaccurate—if a splice is missing or a cable path is mislabeled—the AI will, with high confidence, make a catastrophic error.
The Anatomy of Inaccurate Data: Why Machines Fail
The fundamental challenge is that physical network assets lack inherent intelligence. Unlike a virtualized function or a cloud-native service, a physical fiber optic cable or a street cabinet cannot "self-report." It cannot tell a management system where it resides, how it is connected, or whether it was installed according to the original blueprint.
When human engineers encounter a discrepancy between a schematic and the physical reality, they apply heuristic judgment. They recognize the "smell test" of an error and verify before acting. An autonomous system, however, lacks this intuitive safeguard. It treats the data as an absolute truth.
The consequences of this "trust gap" are multifaceted:
- Flawed Path Calculation: Automated service provisioning can suggest routing that is physically impossible or inefficient because the underlying topology data is incorrect.
- Impact Analysis Failures: During an outage, the system may fail to identify the correct customers affected because the link between the physical asset and the logical service record is broken.
- Field Operations Inefficiency: Dispatching a field technician to the wrong location due to bad data results in wasted operational expenditure and prolonged downtime.
Bridging the Gap: Intelligent Data Transformation
Solving this does not require a "rip and replace" of legacy systems, which is neither financially feasible nor operationally safe. Instead, operators require an "Intelligent Data Migration" strategy that can reconcile legacy, unstructured data into a structured, machine-readable format while the network remains live.
Moving Beyond Manual Redrawing
Historically, the solution to bad data was manual remediation: hiring teams to redraw thousands of legacy files into a new database. This is slow, expensive, and error-prone—often, the same mistakes from the old files are simply transcribed into the new system.
A more effective, modern approach involves Interactively Assisted Converters. These tools utilize AI and machine learning to scan legacy schematics, splice diagrams, and spatial maps, extracting the essential data points automatically. The system handles the heavy lifting of parsing different standards and formats, leaving the human expert to resolve only the "edge cases"—the complex, ambiguous data points that require human judgment.
The Role of Visualization as a Control Layer
As we automate more, the need for human visibility actually increases, not decreases. We require a "Control and Validation Layer" that sits between the automated decision-making engine and the physical reality.
Modern solutions, such as iNTERACTIVE SCHEMATICS™, provide a real-time, visual representation of the network derived directly from the same source of truth the AI uses. By visualizing the data that the machine is seeing, engineers can monitor, audit, and intervene in the autonomous process. This turns the visualization from a static document into a dynamic operational tool that allows for real-time validation of automated decisions.
Strategic Implications: Trust Before Autonomy
The transition to autonomous networks is a strategic imperative, but it is one that requires a shift in priorities. Before an operator can expect to achieve Level 4 or Level 5 autonomy, they must achieve "Data Maturity."
The strategic question every CTO must ask is: Do we have an effective, trustworthy, and scalable way to convert, validate, and reconcile the physical network data on which our future operations depend?
Without this, automation is not an efficiency driver; it is a risk multiplier.
The Roadmap to Readiness
- Audit and Assess: Begin by sampling legacy data to identify the scale of the quality gap. Understanding the variance between as-planned and as-built data is the first step toward remediation.
- Automate the Cleanup: Leverage specialized extraction tools to normalize data across disparate systems.
- Implement a Single Source of Truth: Migrate to a model where the inventory is not just a ledger, but a dynamic, visualized representation of the physical network.
- Human-in-the-Loop Validation: Deploy visualization tools that allow engineers to act as the final arbiter for AI-driven decisions, ensuring that autonomy never moves beyond the guardrails of operational safety.
Conclusion: A Collaborative Path Forward
The path to the future of telecommunications is paved with data. While the industry is rightly focused on the algorithms and AI models that will drive the next generation of connectivity, the true competitive advantage will belong to those who can master their physical data.
At DFG CONSULTING, we believe that the transformation of network data is the most critical hurdle in the current telecom era. We are committed to helping operators convert, validate, and visualize their physical network, turning legacy liabilities into structured assets ready for the autonomous age.
As the industry gathers at events like Connected Britain 2026, the conversation must move beyond the hype of AI toward the pragmatic reality of data infrastructure. We invite operators to reach out for an initial assessment of their data health. After all, if you cannot trust the map, you cannot trust the navigation—no matter how advanced the vehicle.
Miha Ušeničnik has spent more than two decades working across telecommunications, technology development, and operational transformation. Throughout his career, he has held senior technical and management positions in telecom and technology firms, including overseeing broadband network development and the implementation of WiMAX systems. He served as the principal author of Slovenia’s national Broadband Strategy. At DFG CONSULTING, he focuses on the pivotal task of helping network operators transform complex, fragmented network data into trusted, structured information capable of fueling the next generation of automation and AI.
