What challenges did DiMuto face before AI implementation?

Visibility into the packhouse-to-retailer transition was fragmented, causing slow AI model training for fruit defect detection and manual dispute resolution across siloed hybrid systems. Critical data and machine learning workloads were siloed, creating friction for engineering teams and slowing platform reliability and iteration.

What was the impact of the AI solution?

Reduced AI training time by about 35% and reduced costs by about 25%. Improved operational efficiency to accelerate iteration on quality models and agent capabilities.

Source

GCPProductionEvidence: Medium65/100

DiMuto uses Vertex AI and Agent Builder to digitize produce and build trade agents

Use case typeComputer vision inspectionUpdated 1 day ago

DiMuto digitizes every step between packhouse and retailer to track individual produce items and resolve disputes. The company expanded its Google Cloud footprint to run ML and compute workloads, uses Vertex AI for real-time defect detection and grading of scanned cartons, and built a Trade Contract Agent using the Google Agent Development Kit and Vertex AI Agent Builder supported by Gemini models.

Organization: DiMuto
Industry: Consumer & Food
Location: Singapore
Published: August 2026

Reported outcomes

AI training time reduction: Approximately 35% lower

Time & speed

Cost reduction: Approximately 25% lower

Strategic outcomes

Cost efficiencyFaster iteration on quality models and agentsOther strategic outcomeImmutable traceability for trade results

Catalog median for time & speed deployments: −50% across 315 reported metrics. Compare benchmarks →

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

Normalized claim

AI training time reduction: 35% decrease

Google Cloud Customer StoriesAug 1, 2026Customer storyExplicit claimMedium evidence strength

reduced AI training time by ~35%

Normalized claim

Cost reduction: 25% decrease

Google Cloud Customer StoriesAug 1, 2026Customer storyExplicit claimMedium evidence strength

reduced costs by ~25%

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

Customer: DiMuto
Provider: GCP
Maturity: Production
Linked source: Google Cloud Customer Stories

Improved operational efficiency to accelerate iteration on quality models and agent capabilities

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 1 of 1

1Computer vision inspection

Visibility into the packhouse-to-retailer transition was fragmented, causing slow AI model training for fruit defect detection and manual dispute resolution across siloed hybrid systems.
Critical data and machine learning workloads were siloed, creating friction for engineering teams and slowing platform reliability and iteration.

DiMuto expanded its Google Cloud footprint to run ML and compute on Compute Engine and uses Vertex AI for real-time defect detection and grading of scanned cartons.
It built a Trade Contract Agent using the Google Agent Development Kit and Vertex AI Agent Builder supported by Gemini models, and hashes trade results for immutable traceability.

Technologies

Vertex AI Compute Engine Cloud Logging Cloud Monitoring Cloud Run Cloud SQL Cloud IAM Artifact Registry Google Agent Development Kit Vertex AI Agent Builder Gemini

Reduced AI training time by about 35% and reduced costs by about 25%.
Improved operational efficiency to accelerate iteration on quality models and agent capabilities.

Architecture

DiMuto expanded its Google Cloud footprint to run ML and compute workloads. Scanned produce cartons are analyzed by Vertex AI for defect detection and grading, results are hashed onto a public blockchain hosted on Compute Engine for immutable traceability, and a Trade Contract Agent was built with the Google Agent Development Kit and Vertex AI Agent Builder supported by Gemini models.

Sources & evidence1

Evidence: Medium65/100Evidence strength