Project · AI and automation · · Dan Tardif
Google MCP connectors
Building MCP connectors for Google Ads, Google Analytics and Google Tag Manager.
Project objective
I develop connections between my AI assistant and the platforms I use in digital marketing. The goal is to ask a question, retrieve the relevant data and examine it in the same conversation.
The starting point
A marketing question often spans several tools: advertising campaigns, website visits and measurement settings. I wanted to bring that information closer to where I think and work.
MCP is a standard interface that lets an AI application use external tools and data. In this project, each connector exposes specific operations on Google platforms.
Connected platforms
Google Ads
Understand campaigns
Account and campaign inventories, reports and structured diagnostics. Selected actions have a separate human approval process.
Initial live reads verified.
Google Analytics 4
Explore measurement
Read-only access to properties and metadata, with predefined reports covering acquisition, events and trends.
Connection and metadata verified. Live reports still need to be checked.
Google Tag Manager
Review tracking
A connector for inspecting tags, triggers, variables and versions in authorized containers.
Developed and still being brought into service. Writes and publishing are disabled.
Design decisions
Define the scope
The account, property or container must be explicit. Authorized access is part of each tool’s design.
Keep the context
A missing time period, currency or page of results can change an interpretation. Answers retain that context and flag any limitations.
Keep people in control
Reading and changing data follow separate processes. Planned Google Ads actions require approval tied to the specific change.
Connecting the pieces
This project builds on my digital marketing work: start with a practical need, understand the tools and find a more direct way to work with them. AI becomes an interface to operations and data I can examine.
Next steps
Continue checking GA4 reports and bringing Google Tag Manager into service, then document the use cases that become part of my work.