Marketing agency · Australia
Built an AI agent that audits Google Ads daily
An agency was reading raw Google Ads reports by hand to decide what to pause and what to scale. We built them a tool that does the review daily and hands back a ranked list of actions.
The challenge
The team managed several Google Ads accounts and made optimisation decisions by reading platform reports manually. That is slow, easy to get wrong, and wasted spend can sit unnoticed for weeks before anyone catches it. They wanted a system that reviewed performance every day and told them what to act on first, without handing control of the account to an automation.
What we did
- Set up Google Cloud project and Google Ads API access
- Built a secure pipeline pulling campaign, keyword, conversion and cost data
- Structured raw data into decision-ready metrics: CPA, conversion rate and wasted spend
- Built an LLM agent that reviews the data and writes prioritised recommendations
- Defined the optimisation logic for when to pause, scale or change match type
- Kept the agent in advisory mode so the team approves every change