Melbourne Airport has integrated agentic artificial intelligence into its incident response protocols, marking a significant shift in how Australia’s second-busiest airport manages safety-critical events.

The system connects to the airport’s SharePoint document repository, allowing operations staff to retrieve Standard Operating Procedures in real time using natural language queries.

It also cuts out the manual search for compliance documents that has long hampered response times during high-pressure situations.

Irfan Khan, the airport’s Head of Data Analytics, said the problem was straightforward but persistent.

“Critical incident information was historically kept “in people’s heads or SharePoint,” creating unnecessary delays when staff needed to act fast,” he said.

Speaking at the Sydney Microsoft AI Tour, Khan detailed how AI agents now surface the correct SOP for a given scenario within seconds — whether it involves a runway mishap or a medical emergency.

The system can also simultaneously generate a structured incident report for senior leadership, easing the burden on staff who are often fatigued in the aftermath of an incident.

The stakes are considerable. The airport served more than 36 million passengers in 2025 and operates under strict regulatory oversight from Australia’s Civil Aviation Safety Authority for Aerodrome Emergency Planning.

By ensuring staff are always working from the most current approved procedures, the system is designed to strengthen both safety and compliance across one of the country’s most demanding operational environments.

The deployment sits within a broader data strategy built on Microsoft Fabric, which Khan said has already delivered a 30% increase in performance efficiency across data-related tasks.

Looking ahead, the airport expects to extend the technology into predictive maintenance between 2026 and 2027, using data analytics to forecast and prevent equipment failures before they cause disruption.

Data governance, however, remains a live concern. Khan acknowledged the risk of AI agents inadvertently surfacing sensitive personal information linked to employee OneDrive accounts connected to the system.