What challenges did Gelato face before AI implementation?

Engineering support ticket triage required more than 100 hours weekly and only 60% of tickets were correctly assigned. Peak demand created bottlenecks and slowed resolution of tickets. Customer support error categorization involved more than 120 categories and required significant training effort.

What was the impact of the AI solution?

Ticket-triage accuracy increased from 60% to 90%. Gelato saved 120 hours of weekly labor. Time to put ML models into production reduced from weeks to one or two days. Error categorization became faster and more accurate, improving visibility into platform and network performance.

Source

GCP

Gelato used Vertex AI and Gemini to automate engineering ticket triage and error categorization

Use case typeLegal onboarding automationUpdated Jun 13, 2026

Gelato, a Norwegian software company that brings local production to global ecommerce, used Gemini 1.5 Pro on Vertex AI to improve internal engineering-support and customer-support workflows. The company trained Gemini 1.5 Pro to understand its engineering-support ticket-triage process and automatically assign tickets to the correct engineering team. Gelato also embedded Gemini in customer-support systems to instantly classify error reports across more than 120 categories.

Organization: Gelato
Industry: Tech & Comms
Location: Norway
Published: June 2026

Reported outcomes

+90%

ticket triage accuracyQuality & accuracy

60%ticket triage accuracy baseline120 hoursweekly labor saved1.5 daysmodel production time

Strategic outcomes

Speed & agilityAutomated engineering ticket triageNew product / capabilityInstanly classified support error reportsBetter decisions & insightImproved visibility into performance issuesSpeed & agilityAccelerated model deployment to production

Primary read

Use case focus

Showing 3 of 3

1Ticket triage automation
2Customer support automation
3Workflow orchestration

Engineering support ticket triage required more than 100 hours weekly and only 60% of tickets were correctly assigned.
Peak demand created bottlenecks and slowed resolution of tickets.
Customer support error categorization involved more than 120 categories and required significant training effort.

Gelato used Vertex AI to access Gemini models, including Gemini 1.5 Pro.
The team trained Gemini 1.5 Pro on the triage process so reported tickets could be assigned automatically to the right engineering team.
Gemini was embedded into customer-support systems to classify reported errors instantly and support faster resolution.

Technologies

Vertex AI Gemini Gemini 1.5 Pro BigQuery Looker

Ticket-triage accuracy increased from 60% to 90%.
Gelato saved 120 hours of weekly labor.
Time to put ML models into production reduced from weeks to one or two days.
Error categorization became faster and more accurate, improving visibility into platform and network performance.

Sources & evidence1

Groundedness: 5/5Type: Customer StoryPublished: Jun 6, 2026Publisher: Google CloudEvidence: PrimaryConfidence: High

Primary source

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Customer evidence