Join the Sydney Chapter for an evening exploring practical applications of AI in asset management. Across three short presentations, speakers will share how AI, machine learning and computer vision are being applied to complex asset management challenges — from improving system understanding and decision-making under uncertainty, to scaling inspection and condition assessment, strengthening failure classification, and supporting governed, auditable analytics. This event will provide a practical look at how engineering-led AI can support better risk visibility, data quality and asset management decisions.

Building a Practical AI-Assisted Environment for Complex Asset Management Problems

In this presentation, Dr Kevin Hang demonstrates how a structured, engineering led approach to AI, grounded in professional judgement, validation, and practical workflows, can be applied to complex asset management challenges. Focusing on real world implementation rather than theory, he showcases how AI operating within a secure local environment can improve system understanding, enhance risk visibility, and support more informed decision making under uncertainty.

The presentation draws on insights and outcomes from the successful completion of the 2025 Asset Management Council Research Scholarship.

Dr Kevin Hang is an Asset Life Cycle Manager with Sydney Trains, a Chartered Professional Engineer (CPEng) in both Asset Management and Civil Engineering, and a Certified Senior Practitioner in Asset Management (CSAM). He specialises in asset lifecycle optimisation, data driven decision making, and AI assisted analytics for transport infrastructure assets. Dr Hang has developed practical approaches that integrate engineering expertise, machine learning, probabilistic cost-benefit analysis, and local data environments to support infrastructure investment decisions under uncertainty. His work focuses on transforming complex asset, operational, and environmental data into actionable insights that enhance asset performance, resilience, and long-term value.

ML/AI Image Recognition for Asset Inspection and Condition Assessment

This session covers how Ausgrid is applying computer vision and AI to detect and assess defects from inspection imagery, prioritise high-risk assets for early review, and deliver consistent assessments at a scale. Trained on Ausgrid’s defect nuance and built on a reusable cloud-based AI architecture, the capability extends beyond bushfire auditing to support broader asset inspection and condition assessment programs. The same foundation unlocks future use cases including identifying problematic asset inventory, opportunistic analysis of historical photo archives, transforming large volumes of visual data into structured, actionable insights for asset management and network risk.

Lili Chen is a Network Digitisation Technical Lead with extensive experience delivering AI at scale across critical infrastructure, building enterprise-grade AI systems that move well beyond proof-of-concept and into operational use. She specialises in translating innovative AI concepts into reliable, enterprise-scale solutions that support real-world decision-making in operational environments. Combining technical depth with strong domain expertise, her work spans computer vision for asset condition assessment across millions of drone images, satellite and LiDAR-based asset monitoring, predictive vegetation management, vegetation maintenance optimisation, and geospatial AI.

AI Assisted Notification Data Cleaning for Asset Failure Classification and Governance

This session covers how Ausgrid is applying AI, NLP, and rules-based governance to improve the quality and consistency of asset notification data. The capability interprets messy free-text descriptions, Ausgrid-specific abbreviations, and inconsistent failure tagging to recommend more accurate failure classifications, assign confidence scores, and route uncertain cases for SME review. Built with a governed feedback loop between Databricks processing and SharePoint-based SME review, the solution converts AI outputs into maintained rules tables, dictionaries, and failure-tree updates, ensuring the process remains explainable, auditable, and resilient even if AI models change over time. Beyond correcting individual records, the approach helps identify recurring taxonomy issues, failure-tree gaps, and upstream data-capture improvements, creating a scalable foundation for stronger asset analytics, risk modelling, and decision-making.

Xi Chen is a Technical Lead in Asset Modelling at Ausgrid, specialising in asset risk modelling, data-driven decision making, and AI-assisted analytics for electricity network assets. She develops practical approaches that combine engineering knowledge, statistical modelling, NLP and LLM methods, and governed data platforms to improve asset performance, resilience, and long-term investment planning. Her work focuses on turning complex asset, failure, and operational data into scalable, actionable insights for risk modelling, asset management, and regulatory decision support.