Hartsfield-Jackson Atlanta International Airport (ATL) has deployed AI-enabled computer vision overlays across nearly 2,000 cameras, cutting some investigative workflows from four hours to roughly one hour. Chris Crist, the airport's Chief Information Officer, will present the scaling strategy behind this rollout at the International Airport Summit 2026 in Rome from 10-12 November.
The deployment, which Crist's team built on-premises rather than relying exclusively on cloud processing, reflects years of foundational investment in networking, facilities, and operational discipline. "Successful AI initiatives are not solely technology projects. They are organisational change initiatives," Crist said. "Building trust, demonstrating value, and incorporating user feedback are just as important as the underlying technology."
Early results surfaced through investigative workflows. In one case, a query that had required roughly four hours of manual review was completed in about one hour using computer vision-assisted search. Security personnel now use the platform regularly, and additional operational teams are evaluating use cases as familiarity with the tools grows.
Infrastructure before analytics
Supporting real-time computer vision at scale forced the ATL team to confront hardware realities that software-centric pilots rarely surface. GPU-based infrastructure generates significantly more heat and consumes more energy than traditional enterprise technology. Crist's group had to plan hardware placement carefully and upgrade power, cooling, and facility systems before expanding the camera overlay program.
The decision to keep processing within the airport environment, rather than depending on external cloud connections, came from a resiliency requirement. Many use cases involve operational alerting where a connectivity disruption could degrade response times. That architectural choice made infrastructure readiness a prerequisite, not an afterthought.
This pattern repeats across enterprise AI for Operations deployments: the analytics layer may be what users see, but the infrastructure underneath determines whether the system works reliably at scale.
Human judgment stays in the loop
Crist described the objective as providing operational teams with better information, not replacing their judgment. "The goal is to help operational teams investigate incidents more quickly, identify patterns that might otherwise be missed, and make more informed decisions in support of safety, security and efficiency," he said.
Adoption has grown as users discover their own applications. Security teams use the platform for investigations. Other groups are testing it for situational awareness and domain-specific operational challenges. Crist said new use cases continue to emerge as personnel become comfortable with the capabilities.
What comes next
ATL is evaluating computer vision applications across several operational domains: accelerating unattended baggage investigations, detecting unauthorized access to restricted areas, improving curbside activity awareness, and providing earlier visibility into operational disruptions. The airport is also exploring how passenger movement data might inform facility design, concession placement, and amenity decisions.
Crist framed the long-term opportunity around integration rather than any single use case. The aim, he said, is to combine human expertise with real-time intelligence in ways that reduce passenger friction and give operational teams actionable information faster.
Why this matters for operations professionals
ATL's experience surfaces a practical reality that applies well beyond aviation: computer vision at scale is an infrastructure problem first and an analytics problem second. Operations teams evaluating similar deployments should budget for GPU heat loads, power upgrades, and cooling capacity before committing to timelines. The analytics will only be as reliable as the facilities supporting them.
Crist will deliver a full presentation on ATL's deployment at the International Airport Summit in Rome this November.
Your membership also unlocks:
