What challenges did Foodpairing face before AI implementation?

Slow, expensive consumer testing caused product-launch delays and failures. Foodpairing needed to predict consumer preferences for new flavors and packaging faster than traditional surveys and focus groups.

What was the impact of the AI solution?

More than 75% correlation with real consumer panels. Briefing-to-formulation cycle reduced from 18 months to days. Automated flavor analysis of 100 products daily.

Rank#45of 3,117

Source

GCP

Foodpairing uses Gemini Enterprise Agent Platform to predict consumer tastes

Use case typeAutomotive operations automationUpdated Jun 13, 2026

Foodpairing (Belgium) uses Gemini Enterprise Agent Platform and Gemini with BigQuery, Cloud Run, Cloud Storage, Cloud SQL, Firebase, and Google Kubernetes Engine to predict consumer tastes. The company digitizes lab measurements for 100 products per day, streams them into Cloud Storage and BigQuery, and uses roughly 200,000 digital twins plus Gemini-powered agents to simulate reactions to new flavors and packaging. The article claims more than 75% correlation with real consumer panels and says the briefing-to-formulation cycle dropped from 18 months to days.

Organization: Foodpairing
Industry: Consumer & Food
Location: Belgium
Published: June 2026

Reported outcomes

+75%

consumer panel correlationQuality & accuracy

−99.7%briefing-to-formulation cycle time100 products/dayproducts analyzed daily

Strategic outcomes

New product / capabilityBuilt virtual consumer simulation platformSpeed & agilityAccelerated briefing-to-formulation cycleBetter decisions & insightImproved prediction of consumer preferencesScale & capacityAutomated large-scale daily flavor analysis

Catalog median for quality & accuracy deployments: +90% across 287 reported metrics. Compare benchmarks →

Primary read

Use case focus

Showing 3 of 3

1Agentic AI
2Digital Twin Simulation
3Predictive Analytics

Slow, expensive consumer testing caused product-launch delays and failures.
Foodpairing needed to predict consumer preferences for new flavors and packaging faster than traditional surveys and focus groups.

Built a platform on Google Cloud with BigQuery, Cloud Storage, Cloud Run, Cloud SQL, Firebase, and GKE.
Used Gemini Enterprise Agent Platform and Gemini to model virtual consumers and simulate market launches.
Automated analysis of 100 products daily and ran simulations on Cloud Run to test dozens of concepts in a morning.

Technologies

Gemini Enterprise Agent Platform Gemini BigQuery Cloud Run Cloud Storage Cloud SQL Firebase Google Kubernetes Engine

More than 75% correlation with real consumer panels.
Briefing-to-formulation cycle reduced from 18 months to days.
Automated flavor analysis of 100 products daily.

Architecture

Foodpairing built its platform on Google Cloud, using Cloud Storage and automated pipelines into BigQuery to ingest daily lab measurements, Cloud Run to scale simulations, Firebase and Cloud SQL for the Inspire chef platform, and Google Kubernetes Engine as part of the broader stack. Gemini Enterprise Agent Platform and Gemini power agents that simulate reactions of roughly 200,000 digital twins and communicate to approximate consumer market launches.

Sources & evidence1

Groundedness: 5/5Type: Customer StoryPublished: Jun 13, 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