If you’re running paid advertising, whether that’s a lean Meta budget for a local business or a multi-client media operation, there are two AI tools that have changed our team’s day-to-day and how we operate. Not in a press-release kind of way, but in a “this saved me an hour on a Monday when I had six meetings” kind of way.
Here’s how we actually use them.
Claude for copy, reporting, and everything in between.
Claude by Anthropic has become the connective tissue of our paid media workflow. Claude is an AI assistant, but the range of things we use it for on any given day is wide.
Copy & Headline Writing:
If you’re running Meta ads and you have one piece of primary text you’re reasonably happy with, Claude can generate fifteen headlines and copy variations from it in minutes. Some won’t be usable, truly most won’t be. But two or three of them could end up being better than what you started with, and now you have a real creative testing set instead of minor rewrites of the same idea. The variations that perform best in-platform tell you something. You put more dollars behind those.
For Reporting:
Download your campaign performance data (say, three months of metrics across all your active and now inactive ads), upload it to Claude, and ask it to analyze your performance – specifically what’s working, what isn’t, and where shifting spend might improve results. What would take an analyst an hour or more to manually look through and summarize comes back in a few minutes. It’s not a replacement for human judgment, but it surfaces the right questions faster so you can get to the actual decision-making sooner.
Other Uses:
Beyond paid media, Claude is useful for daily workflow management in ways that compound over time. A practical example: on a recent Monday, we had six back-to-back meetings. Rather than spending ten to fifteen minutes preparing for each one, we asked Claude to pull the relevant context from the email and previous meetings, build a single briefing sheet covering all six, and flag anything that needed action. What would have taken over an hour across the day took a few minutes in the morning.
Getting to the point where Claude consistently understands how you work takes time. It took over a week of regular back-and-forth to train it to produce outputs that felt right for our team. That’s not a complaint, but more the reality of any AI tool worth using. The investment pays off, but it does take some time and effort to fine-tune Claude to your expectations.
It’s not a fix-all solution, but rather a tool at the end of the day.
Meta AI for creative variation and in-platform optimization.
Meta’s native AI features live inside Ads Manager, which means most people running ads already have access to them and may not be using them intentionally or at all.
A few worth knowing:
Copy Suggestions:
Write your own primary text and headline and let Meta suggest alternatives. Not all of them will be production-worthy, but having five options on the table instead of one changes the conversation about what to test. And from time to time, the AI-generated version may even outperform the one you wrote.
Creative generation:
Upload a static image and Meta will generate variations, including different crops, color treatments, and compositions based on the original content. This is most useful for broad campaigns such as seasonal promotions, event announcements, and awareness campaigns. It’s less useful when you’re advertising a specific product where the photography is the entire focus. AI-generated images of a generic “Mother’s Day” concept work fine, while AI-generated images trying to replace your signature product usually don’t work.
Budget and bidding suggestions.
Meta constantly notifies you of these suggestions mid-campaign. You don’t have to follow them, but they’re worth reading rather than dismissing. Sometimes the suggestion catches something you wouldn’t have noticed on your own.
Neither tool replaces the human in the loop. Claude will occasionally miss the tone you’re going for, while Meta will generate images with unreadable text or awkward compositions. The value isn’t in taking the output and running with it, but in having a high volume of options to react to, edit, and improve, in a fraction of the time it would take to produce them from scratch.
That’s the shift worth internalizing. AI doesn’t do the job for you; it means you spend less time on the parts that don’t require your expertise, and more time on the parts that do.
