This repository provides a comprehensive solution for leveraging Google Cloud's generative AI capabilities in marketing. It targets marketers and developers seeking to automate content creation, personalize campaigns, and gain data-driven insights, offering a deployable application with detailed guidance and supplementary notebooks.

How It Works

The solution integrates Google Cloud's Vertex AI for generative AI tasks, including text and image generation, and Vertex AI Search for content discovery. It features a conversational interface for natural language to SQL queries (nl2sql) via Looker dashboards for marketing insights and campaign analysis. Content generation is driven by configurable prompts and foundation models, with options for fine-tuning and integration into Google Workspace for seamless asset transfer.

Quick Start & Requirements

• Deployment: Follow instructions in the deployment guide using Terraform.
• Prerequisites: Google Cloud account, Terraform, Python. Specific Vertex AI models and Looker access are required.
• Resources: Deployment involves setting up various Google Cloud services.
• Links: Deployment Video, Notebooks

Highlighted Details

• Demonstrations: Includes Marketing Insights (Looker dashboards), Audience/Insight Finder (nl2sql), Trendspotting (Google Trends), Content Search (Vertex AI Search), Content Generation (text/visuals), and Workspace Integration.
• Notebooks: Covers nl2sql, news summarization with LangChain/Vertex AI, Imagen fine-tuning, Gemini tuning, document Q&A, and Vertex AI Search examples.
• Configuration: Managed via infra/variables.tf and config.toml for project settings, prompts, model selection, and data personalization.
• Customization: Allows embedding custom Looker dashboards into the frontend UI.

Maintenance & Community

• Support: Report issues via GitHub issues.
• Links: GitHub Issues

Licensing & Compatibility

• License: Apache License 2.0.
• Compatibility: Permissive license suitable for commercial use and integration with closed-source applications.

Limitations & Caveats

The solution requires significant Google Cloud resource configuration and familiarity with Terraform for deployment. Customization of Looker dashboards involves direct HTML modification within the frontend code.