An AI marketing department in a Terraform module

What it does

This repository is a full-stack reference solution that wires Google Cloud services into a marketing toolkit. It gives marketers a web interface to query BigQuery data via natural-language-to-SQL, spot trends using Google Trends, search content with Vertex AI Search, and generate copy and images with Vertex AI models. The entire stack—frontend, backend APIs, and infrastructure—is packaged for deployment via Terraform.

The interesting bit

Rather than leaving generated assets in a chat window, the solution pushes emails, blog posts, and ad copy directly into Google Workspace apps like Slides, Docs, and Sheets. It also tries to eliminate the BI bottleneck by letting marketers ask plain-English questions that get translated into GoogleSQL queries against their own data.

Key highlights

• Deployable as a unit via Terraform, including frontend, backend, and infrastructure scripts.
• Conversational NL2SQL interface for BigQuery marketing data.
• Trendspotting pipeline combining Google Trends analysis with LangChain-based news summarization.
• Content generation for text and images using Vertex AI foundation models, plus notebooks for Imagen fine-tuning.
• Native Google Workspace integration to export and visualize outputs.
• Supplementary Jupyter notebooks covering ReAct agents, document Q&A, and Gemini tuning.

Caveats

• Deeply embedded in the Google Cloud ecosystem; meaningful use requires Vertex AI, BigQuery, Looker, and Workspace.
• Adding custom Looker dashboards requires hand-editing Angular HTML component files, which is an awkward fit for a “no-code” marketing audience.
• The extensive deployment and configuration guidance suggests this is a proof-of-concept showcase rather than a polished product.

Verdict

A useful starting point for GCP customers who want a working demo to pitch AI-driven marketing to stakeholders. Everyone else will likely find the platform coupling and setup overhead hard to justify.