What challenges did CrossTech face before AI implementation?

Traditional rail and road inspections are time-consuming and costly. CrossTech needed to process large volumes of video and lidar-derived inspection data quickly to identify hazards proactively and reduce unplanned service interruptions. The company also needed scalable infrastructure that supported its AI platform and deployment speed while controlling costs.

What was the impact of the AI solution?

50% reduction in time to build AI models. 70% reduction in time to deploy new models and features. 96% reduction in high-risk faults on key lines. 30% reduction in unplanned service interruptions. Approximately £20M per annum in net maintenance efficiency. Frontline teams remediated 6,000 faults.

Source

GCPEvidence: Medium50/100

CrossTech: predictive maintenance and computer vision inspection with Vertex AI

Use case typeWind farm maintenanceUpdated Jun 13, 2026

CrossTech, a UK transport network inspection company, built an automated AI infrastructure inspection platform called Hubble to analyze video data captured from trains and identify hazards such as overgrown vegetation, signal obstructions, level crossing sighting risks, and track ballast issues. The company migrated to Google Cloud and built a containerized microservices architecture using Cloud Run and Compute Engine for near-real-time processing and autoscaling, with Vertex AI and Vertex AI Notebooks used to accelerate model development and prototyping. The article also notes continued experimentation with Gemini Enterprise to automate tasks and streamline internal agentic workflows, and with Firebase Studio to simplify coding.

Organization: CrossTech
Industry: Logistics
Location: United Kingdom
Published: May 2026

Reported outcomes

−96%

quantified impactRisk, reliability & safety

−50%time−70%time−30%quantified impact

Strategic outcomes

New product / capabilityBuilt an automated AI inspection platformSpeed & agilityAccelerated model development and prototypingSpeed & agilityStreamlined and automated deploymentsScale & capacityEnabled near-real-time inspection processing

Catalog median for risk, reliability & safety deployments: −70% across 13 reported metrics. Compare benchmarks →

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Time: 50% decrease

Google Cloud Customer StoryMay 24, 2026Customer storyInferred claimMedium evidence strength

50% reduction in time to build AI models.

Normalized claim

Time: 70% decrease

Google Cloud Customer StoryMay 24, 2026Customer storyInferred claimMedium evidence strength

70% reduction in time to deploy new models and features.

Normalized claim

Quantified impact: 96% decrease

Google Cloud Customer StoryMay 24, 2026Customer storyInferred claimMedium evidence strength

96% reduction in high-risk faults on key lines.

Normalized claim

Quantified impact: 30% decrease

Google Cloud Customer StoryMay 24, 2026Customer storyInferred claimMedium evidence strength

30% reduction in unplanned service interruptions.

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: CrossTech
Provider: GCP
Maturity: Unknown
Linked source: Google Cloud Customer Story

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 4

1Predictive Maintenance
2Computer Vision
3Infrastructure Inspection

Traditional rail and road inspections are time-consuming and costly.
CrossTech needed to process large volumes of video and lidar-derived inspection data quickly to identify hazards proactively and reduce unplanned service interruptions.
The company also needed scalable infrastructure that supported its AI platform and deployment speed while controlling costs.

CrossTech built a containerized microservices computer vision pipeline on Google Cloud.
Cloud Run and Compute Engine autoscaling support near-real-time processing and cost control.
Vertex AI and Vertex AI Notebooks speed up model development lifecycle and prototyping.
App Engine is used to streamline and automate deployments.

Technologies

Vertex AI Vertex AI Notebooks Cloud Run Compute Engine App Engine Gemini Enterprise Firebase Studio

50% reduction in time to build AI models.
70% reduction in time to deploy new models and features.
96% reduction in high-risk faults on key lines.
30% reduction in unplanned service interruptions.
Approximately £20M per annum in net maintenance efficiency.
Frontline teams remediated 6,000 faults.

Architecture

CrossTech uses a containerized microservices architecture on Cloud Run and Compute Engine for its computer vision pipeline. Vertex AI and Vertex AI Notebooks support faster model development and prototyping, while App Engine helps automate deployments. The company is also experimenting with Gemini Enterprise and Firebase Studio for internal workflow and development assistance.

Sources & evidence1

Evidence: Medium50/100Evidence strength