Why I Think AI Optimization Tools Aren’t Quite Ready Yet

AI-powered optimization tools hold great potential, and they might become highly effective in the future, but they are not ready yet. For now, traditional software and manual account managers still offer a more reliable approach, especially when it comes to account optimization in the current AI era.

Let me share a quick example we came across during an account audit. This particular account was entirely managed by an AI software agent, without any human intervention.

In SD campaigns, the AI was creating a separate ad group for each product line. As a result, every SD campaign ended up with over 200 ad groups containing different products. Within each of these ad groups, the AI was adding product targeting based on search terms pulled from SP and SB campaigns.

Unfortunately, most of these product targets were either completely irrelevant, often books or unrelated items, or poorly matched. Worse, the AI kept adding new product targets daily based on STRs, leading to significant inefficiencies. On average, SD campaigns showed about 50% wasted spend due to these irrelevant targetings. A similar pattern was seen in SP and SB campaigns.

On top of that, the bid optimization logic was superficial. It adjusted bids daily or even multiple times a day but with negligible changes, such as increasing a bid from $0.06 to $0.07 or $0.21 to $0.22. These minor tweaks did not have any real impact on performance. As a result, keyword performance was either stagnant or declining.

Negative keyword targeting was also mismanaged. The AI system was adding even relevant keywords to the negative list on a daily basis, only to archive them later. It failed to distinguish between relevant, generic, or non-performing keywords, lumping everything together indiscriminately.

If you're using AI optimization tools, be cautious and monitor your campaigns closely. You do not want to end up wasting ad spend on irrelevant or poorly targeted placements.