Unilever · Google Cloud
15-35% increase — retailer sales
Independently reportedOther · Other · Classical ML · Scaled
Other
Day of AI Australia / UNSW Sydney · Google Cloud
Vendor-reportedEducationAustraliaIn productionPublic sectorLLM
Vendor-reported. The customer is named and the numbers are quoted from the source page, but the account comes from the vendor. No independent confirmation.
Reported by google.com (vendor self-report). Checked against the source page on 2026-09-06. 6 of 6 figures below appear on that page word for word.
The problem
Moving AI education from passive observation to active participation, processing thousands of unique student-generated prompts in real-time, and ensuring 100% safety compliance for generative AI content created by 12-16 year olds.
What was deployed
- Vendor
- Google Cloud
- Products
- Compute Engine, Shield Gemma, Model Garden on Vertex AI, Vertex AI, Cloud SQL, Gemini in Vertex AI
- Models
- Gemma, Gemini
- Technique
- LLM
- Build or buy
- Bought and customised
- Deployment
- SaaS
- Data used
- student-generated prompts, encrypted game states and teacher logs
- Scale
- over 330,000 students, 350+ concurrent student teams, 120,000+ social posts
- Timeline
- 3 months
What changed
Each row is quoted from the source. Figures we could not find on the page in those words are marked — they are kept, not deleted, so you can judge them.
330000Students supportedto date
“Scales safe AI simulations to support over 330,000 students to date with Compute Engine”
Quoted word for word from the source
100Harmful content blocked
“Blocks 100% of harmful content using Shield Gemma”
Quoted word for word from the source
120000Safe AI social posts generated
“Generates 120,000+ safe AI social posts for "Win the Farm" using open Gemma models”
Quoted word for word from the source
350Concurrent student teams supported
“Supports 350+ concurrent student teams with auto-scaling from Vertex AI”
Quoted word for word from the source
0Performance issues
“The platform powered the intensive "Win the Farm" competition without a single performance issue.”
Quoted word for word from the source
0Safety incidents
“It generated more than 120,000 unique social media posts without a safety incident.”
Quoted word for word from the source
“A Year 8 student admitted they never realized content on TikTok could be fake until they built a bot to do it themselves.”
Difficulties and limits
- Giving 12 to 16-year-olds the ability to direct a generative AI model posed a considerable risk (generating hate speech or age-inappropriate content)
- Relying on fragmented tools from multiple providers for billing, inference, and security monitoring proved inefficient for this lean non-profit team
Sources
- google.com — vendor self-report
Similar deployments
ANZ Bank · NVIDIA
0.82 — Gini coefficient for assessing risk
Independently reportedBanking & finance · Other · Classical ML · Pilot
BHP Billiton
zero — accidents due to driver drowsiness
Independently reportedOther · Other · Classical ML · In production · 2022
DSST: College View · Playlab, Class Companion, Manta AI
Vendor-reportedEducation · Other · NLP · In production · 2024
Cubo Ai · Google Cloud
more than 10X — user growth supported with same IT workforce
Vendor-reportedOther · Other · Computer vision · Scaled · 2019
Precision HVAC & Cooling · AutoScale My Biz
15+ — High-Value After-Hours Leads Captured
Vendor-reportedOther · Other · LLM · In production