← All projects
Internet marketplace / AI automation

Five AI agents to reduce paid search waste.

I built agents to identify bidding inefficiencies across a $2.3M advertising budget, with recommendations representing an estimated $309K in annual savings.

$309K

Estimated annual savings

The problem

An internet marketplace needed to identify waste in its paid search spending. Reviewing bidding inefficiencies was a recurring analytical task across a $2.3M budget.

What I built

I built five AI agents to identify bidding inefficiencies and support corrective recommendations. The work combined automation with analysis: finding where spend could be reduced and explaining the commercial implications to executive stakeholders.

What the result means

The recommendations represented approximately $309,000 in estimated annual savings. That is an annualized estimate, not a claim that a full year of savings had already been realized.

What I took from the work

Automating analysis is useful when someone can act on the output. I had to structure an unfamiliar problem, connect the findings to advertising spend, and communicate a recommendation leadership could use.