Original article Total Telecom:Read More
Contributed Article
by Miha Ušeničnik, Associate Director at DFG CONSULTING
Telecom automation is moving beyond predefined tasks towards networks that can increasingly analyse conditions, make decisions and act autonomously. Self-configuration for provisioning, closed-loop optimisation, and automated fault recovery promise greater operational efficiency and reliability, while enabling services to be provisioned and adapted with less manual intervention.
The transition is already underway. According to a study by the IBM Institute for Business Value and TM Forum: Navigating autonomous networks, 73% of network executives surveyed said their organisations had developed phased roadmaps towards autonomous operations. Yet only 6% of CSPs reported operating highly autonomous Level 4 network instances. Within three years, 22% expect to reach that level.
What must change to bridge that gap? It starts with data. As a data transformation company, DFG CONSULTING focuses on the quality and trustworthiness of the physical network data that automation depends on.
When the network and its data diverge
Telecom networks have evolved continuously over decades, but the data describing them has not always kept pace. Network extensions, technology changes, mergers and system migrations have left many operators with information distributed across multiple systems and formats.
Critical network information may still reside in CAD or Visio drawings, PDFs, raster images and spreadsheets alongside central inventory systems. While valuable to engineers, this unstructured information cannot readily support automation.
The challenge goes deeper than format. Field changes are not always accurately reflected in the system of record, creating discrepancies between as-planned and as-built networks.
Missing or incorrect connectivity creates further uncertainty. Engineers and field teams may consequently maintain local “shadow documentation” to capture information that is missing, outdated or difficult to retrieve from the system of record.
The problem is therefore often not the absence of data, but whether operators can trust that it accurately represents the physical network and whether it is in the right form for growing network automation and autonomy.
The consequences of inaccurate data
Experienced engineers might still recognise when a network record does not reflect reality and question it before acting. Automated operations do not necessarily have that safeguard because they cannot recognise inaccurate data.
Physical assets do not have intelligence, telemetry, or at least some form of active communication. They cannot report their state. They cannot self-discover. They cannot tell an inventory system where they are, how they’re connected, or whether they were built to design. Automated decisions solely rely on a trustworthy, precise, end-to-end digital representation of all available physical assets.
Incorrect data can lead to an incorrect path calculation, affecting service provisioning or instructions for field teams. Inaccurate connectivity can result in flawed impact analysis and the wrong customers or services being identified during an outage.
Intelligent data migration
Recognising the need for trustworthy data is the easy part — the challenge is quality, scale, efficiency and continuity. Operators may hold hundreds of thousands of legacy files accumulated over decades, while new documentation keeps arriving, often incomplete, inconsistent or conflicting, and daily operations cannot pause while this data is transformed.
The question becomes: how can operators convert, validate, and reconcile network data at scale while updating operational data on the fly?
Physical network documentation is naturally fragmented across distinct views — spatial maps, schematics and splice diagrams — each describing the same elements at a different level of detail. Manually redrawing legacy files into operational support systems is common, but a more effective approach is to use a specialised tool Interactively Assisted Converter™ that extracts data from legacy drawings and structures it in a machine-readable format, ready for direct ingestion into the system of record.
This tool can cross-reference and merge disparate views into a single dataset, with AI adapting to the multiple drawing standards different data providers use. Data conversion is fundamentally a cleaning process: automation extracts, structures, and resolves quality issues for most of the data, while human oversight handles conflicting edge cases — far more reliable than manual redrawing, which is resource-intensive and prone to replicating existing mistakes into the central system of record.
Keeping humans in control
As autonomous systems make decisions at a scale and speed humans cannot replicate, engineers may struggle to understand and independently validate them. Human-readable network connectivity visualisation therefore becomes more important, not less: data transformation gives us machine-operational data, but at the cost of human readability.
The answer isn’t another manually maintained layer of “shadow” documentation. Instead, iNTERACTIVE SCHEMATICS™ can automatically generate high-level and low-level network diagrams directly from the same inventory data used for automation — providing a real-time visual representation from the single source of truth.
Network connectivity visualisation is now evolving from documentation into a control and validation layer between automated systems and engineers, turning trusted digital data into an interactive operational view to support design, planning, troubleshooting, maintenance, and service provisioning.
Trust before autonomy
Autonomous networks are not only about making machines more intelligent. They also require a trustworthy representation of the physical network and the human visibility needed to understand and validate manual and automated decisions.
As network operators move towards higher levels of autonomy, the strategic question is:
Do we have an effective, trustworthy and scalable way to convert, validate, reconcile and visualise the physical network data on which network operations depend?
DFG CONSULTING can help you answer that question.
We can assess a representative sample of your own data to identify quality gaps, transformation opportunities and potential process improvements — and determine what it takes to make your network data trusted and automation-ready.
Contact miha.usenicnik@dfgcon.si to arrange an initial assessment or meet us at Connected Britain 2026 on 9–10 September.
Miha Ušeničnik has spent more than two decades working across telecommunications, technology development and operational transformation.
He has held senior technical and management positions in telecom and technology companies, including responsibility for broadband network development and WiMAX systems, and contributed to national broadband policy as principal author of Slovenia’s Broadband Strategy.
At DFG CONSULTING, his focus is on helping network operators transform complex network data into trusted, structured information that can support more efficient operations, automation and AI.
The post Garbage in, bad decisions out: The data problem behind autonomous networks appeared first on Total Telecom.