Mexican airports move 133M passengers. No vendor has named one.

Synaptic at Ezeiza is the LATAM bar. Pearson at 44% taxi-in is the global. GAP, ASUR, and OMA are the white-space.

Proven pattern

Camera-based computer vision detecting and timestamping every event in a turnaround (chocks on/off, GPU connection, jet-bridge attach, fueling start/end, catering, APU dwell, baggage start/end, pushback ready), feeding an analytics layer that benchmarks performance, flags anomalies, and supports OTP and emissions reporting.

Hard KPIs · operator-disclosed where possible

44%

taxi-in time reduction (Pearson/Assaia, vendor-disclosed)

all gates

Synaptic at Ezeiza after 90-day POC (Feb 2026)

~$100/min

block-time cost on the ramp (FAA)

Buyer · COO of the airport authority/operator + airline operations counterpart. The airport is the prime; the airlines are the data co-providers.

Operator references

Provenance tiered explicitly. Audited > Operator-disclosed > Vendor-published. Treat all vendor-tier figures as directional.

Ezeiza Buenos Aires (AA2000) · Synaptic Aviation

AI ramp-visibility platform deployed across all gates (Feb 2026, after a 90-day POC). >60 cameras. Computer vision detects chocks, GPU connection, jet-bridge attach, fueling start, catering, APU dwell.

Operator-disclosed

Toronto Pearson · Assaia (ACI-recognized)

Vendor-published 44% reduction in average taxi-in time and CAD 47M annual savings. Treat as directional.

Vendor-published

GRU São Paulo

ARINC VeriPax passenger-reconciliation system claimed at 15–20% passenger-flow improvement. Vendor-aligned, directional.

Vendor-published

Delta APEX (US, airline-side)

Predictive maintenance reduced maintenance-related cancellations from 5,600/year to 55. Eight-figure annual savings; 2024 Aviation Week Grand Laureate. Adjacent: same data architecture, airline-owned.

Audited

GAP / ASUR / OMA (Mexico) · explicit absence

No public, named AI vendor turnaround deployment for 2023–2026 across collectively 133M+ PAX. Capital is going to capacity expansion (GAP MX$52B through 2029; OMA MX$8B for Monterrey 2026–2030) rather than operational AI.

Audited

Why agents and CV beat status-quo manual logging

Status quo is gate agents on radios and paper. The data does not exist at the granularity needed for systematic improvement. CV produces structured per-event data; the agent layer reasons about cross-stand patterns, anomalies, and corrective recommendations.

Why this is hard

Camera infrastructure and connectivity. Multi-stakeholder data sharing (airport ↔ airline ↔ handler). Procurement cycles in LATAM are 18–36 months. Concession-model fragmentation in Mexico (GAP/ASUR/OMA all separate publicly listed groups).

Mexico · LATAM specifics

GAP runs Guadalajara, Tijuana, Los Cabos, Puerto Vallarta + 9 more; ASUR runs Cancún + 8 more in the southeast; OMA runs Monterrey + 12 more in the central/north. Cancún (ASUR) is the highest-traffic LATAM tourism hub and the most strategically valuable single deployment surface. Guadalajara and Monterrey are the cleanest secondary-hub pilot candidates.

Reference architecture

Each node maps to a regulatory anchor.

01
Camera infrastructure: fixed cameras + perimeter coverage
02
CV detection layer: per-event with confidence and source frame
03
Event stream: normalized schema across gates and stands
04
Anomaly + analytics agent: cross-stand reasoning
05
AOC dashboard: operations + sustainability + OTP
06
Airline / handler API: explicit data-sharing boundary

Anti-positioning

Not A drone tour. A futuristic-airport CGI. A tech-conference reel.

But Wide ramp shots at architectural daylight. Closer to Synaptic's Ezeiza coverage than to any vendor reel.

What didn't work initially

Early deployments fail when the airport tries to consume CV data without a named data-sharing agreement with the airlines. The fix is a 90-day POC at one stand with one airline partner, before the airport-wide rollout.

A working session, not a sales call.

Two hours with a partner. Your incumbent stack, your data posture, and the regulatory surface against a sovereign reference architecture for aviation.

Request a working session →

A working session, not a sales call.

Two hours with a partner. We map your AI spend, data exposure, and governance posture against a sovereign reference architecture. You leave with a memo. We leave with a decision.