Telstra: AI Workloads Expose Data Gaps in Telecom Automation

AI demand reshapes network operations

AI-driven traffic is forcing telecommunications operators to rethink not just capacity, but the fundamental automation strategies that keep networks running. According to Regan Ireland, global head of pre-sales solutions, products and digital experiences at Telstra, the past 12 to 18 months have brought unprecedented scale and unpredictability to network operations.

Speaking at Fierce Network's AI and the Automated Network virtual event, Ireland described growth from hyperscalers and emerging neoclouds as "unfathomable," with demand shifting rapidly in both magnitude and geographic location. That volatility is raising the bar on how quickly operators must respond and how reliably they can automate across their own infrastructure and partner networks.

Why it matters

As AI workloads become a dominant traffic source, operators face a strategic fork: automate faster to meet demand, or risk service degradation and customer churn. But the path to autonomy isn't a simple technology upgrade — it requires resolving longstanding data quality issues and deciding where human judgment remains essential. Telstra's approach offers a pragmatic model for balancing speed with accountability.

Autonomy must be earned, not just built

Telstra is accelerating its move toward autonomous operations, but Ireland emphasized that autonomy is "something you have to build, but it's also something you have to earn." The operator is prioritizing automation in lower-risk, repeatable tasks while keeping humans accountable for policy decisions and risk management.

Before operators can trust automated systems at scale, Ireland said they need "clean telemetry," strong observability, and a "governed source of truth." Today, Telstra sits somewhere between traditional automation and full autonomy, using automated systems to speed detection and recovery while preserving human oversight for high-impact decisions.

Engineers still determine policy, manage exceptions, and evaluate risks tied to customer impact. "Anything to do with policy, with complex exceptions…those sorts of complex trade-offs, they are something that I think there will always be a human in the loop," Ireland said. What Telstra aims to eliminate is operational drag — the manual handoffs, reconciliation steps, and redundant checks that slow response times.

Data quality remains the critical bottleneck

The industry's biggest constraint remains data. Ireland said data "still remains the biggest challenge," particularly around telemetry. The issue isn't collecting more data, but making it usable through normalization and correlation. Without that, "the signal just gets lost in the noise," he said, describing how operators can drown in alarms without extracting actionable meaning.

Visibility gaps across network boundaries compound the problem, forcing continued reliance on manual validation where systems don't align. These data challenges directly limit how far operators can push automation without introducing new risks.

Customer experience as the automation metric

For customers, increased automation should translate to proactive operations. Ireland pointed to a shift away from "we'll get back to you" responses toward faster quotes, better visibility into service delivery, and more consistent incident handling. The goal is a model where customers can see service status in real time and understand risks and mitigations as they occur.

The end state isn't a hands-off network, Ireland made clear. It's one where automation removes repetitive work while human judgment continues to define how operators manage risk, resilience, and customer impact.

These details were first reported by Fierce Network during its AI and the Automated Network virtual event.