-
I’ve gone a bit down the rabbit hole with this tbh. At this point I’m not really using AI as a “tool” anymore, it’s more like how I operate accounts.
I built a system around search terms that does more than just negatives. It pulls data, groups queries, flags irrelevant patterns, and surfaces gaps in coverage. Not perfect, but it saves a stupid amount of time and forces consistency. Before it was very manual and easy to miss things.
On the reporting/analysis side, I’ve stopped building dashboards the traditional way. I just pull structured data and run analysis on top of it. Things like “what actually changed week over week and why” instead of just showing numbers. That’s probably the biggest shift for me.
I’ve also built onboarding flows where I can take a new client, run keyword research, cluster it, and get to a first campaign structure pretty quickly. Still needs human input obviously, but it removes a lot of the repetitive groundwork.
Outside of the Ads themselves, I’ve probably gotten the most leverage from just connecting things. If there’s an API, I’ll try to plug into it. CRM, forms, tracking, whatever. Then AI sits on top and helps make sense of it or trigger actions. That part is honestly more valuable than anything inside the ad platform.
Biggest lesson for me though is that most “AI PPC tools” are underwhelming because they lack context. If you don’t control the inputs and structure, the outputs are average at best.
Once you start building your own stuff around your workflows, it becomes a lot more useful. Each of my clients has their own “context folder” on my device. This automatically pulls in emails, their website, meeting transcripts etc. which I can then leverage for much more comprehensive data analysis and decision making for a client.
So I haven’t really vibed coded a smart app or software, I’ve just built a system that leverages the available API’s from relevant ad accounts, that makes it possible to make smarter decisions for my clients.
