What challenges did Overdose Digital face before AI implementation?

Reduce dashboard fatigue and the need for manual interpretation across hundreds of clients. Overcome enterprise trust concerns about hallucinations and untrusted AI outputs. Move from static reporting to an automated insight-to-action creative production loop.

What was the impact of the AI solution?

Near-zero hallucinations through grounding in Looker semantic definitions. Automated creative production and optimization feedback loops while maintaining brand safety. Enabled conversational analytics and faster action from data.

Source

GCP

Overdose Digital: Agentic creative production grounded with Looker + Vertex AI/Gemini

Use case typeAI platformUpdated Jun 13, 2026

Overdose Digital, an AI-first consultancy and digital commerce agency, centralized ecommerce data in BigQuery and used Looker as a governed semantic layer to ground generative AI and reduce hallucinations. The company built an agentic workflow where Looker-defined metrics trigger Vertex AI and Gemini-based agents to generate hyper-localized creative briefs and assets with human oversight. It also used BigQuery AI to detect creative fatigue and automatically trigger a synthesis loop that refreshes creative strategy near real time for nearly 300 clients.

Organization: Overdose Digital
Industry: Retail
Location: Australia
Published: June 2026

Reported outcomes

Strategic outcomes

Risk & complianceReduced hallucinations in AI outputsNew product / capabilityAutomated creative production workflowBetter decisions & insightEnabled conversational analyticsSpeed & agilityRefreshed creative strategy near real time

Primary read

Use case focus

Showing 3 of 4

1Generative AI for marketing and creative production
2Conversational analytics
3Agentic workflow orchestration

Reduce dashboard fatigue and the need for manual interpretation across hundreds of clients.
Overcome enterprise trust concerns about hallucinations and untrusted AI outputs.
Move from static reporting to an automated insight-to-action creative production loop.

Centralized client ecommerce data into BigQuery.
Used Looker as the semantic brain/source of truth to define metrics and power conversational analytics.
Implemented an event-bus pattern where Looker detects KPI patterns and triggers Vertex AI agents, including Gemini, to generate localized creative assets grounded in Cloud Storage brand assets.
Used BigQuery AI to detect creative fatigue and automatically refine and refresh assets near real time.

Technologies

Looker BigQuery Vertex AI Gemini Cloud Storage Imagen Veo

Near-zero hallucinations through grounding in Looker semantic definitions.
Automated creative production and optimization feedback loops while maintaining brand safety.
Enabled conversational analytics and faster action from data.

Architecture

Overdose Digital centralized ecommerce data into BigQuery, used Looker as the governed semantic layer and intelligence layer, then triggered Vertex AI/Gemini-based agents through an event-bus pattern when KPI conditions were detected. Brand assets in Cloud Storage grounded asset generation, and BigQuery AI detected creative fatigue to trigger a synthesis loop that refreshed creative strategy and assets.

Sources & evidence1

Groundedness: 5/5Type: Customer StoryPublished: Jun 7, 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