Last month, while auditing one of the accounts, we discovered many missed opportunities. The account was entirely managed by AI, with no human intervention, and we found a few important issues such as:

- AI created 200+ ad groups per SD campaign, each tied to a product line, with product targets pulled from SP/SB search terms.
- Most product targets were irrelevant (e.g., books or unrelated items), with daily additions from STRs leading to approximately 50% wasted spend across SD, SP, and SB campaigns.
- Bid optimization was ineffective, with frequent micro-adjustments (e.g., $0.06 to $0.07) that had no real impact on performance.
- Negative keyword logic was flawed, often blocking relevant terms and failing to differentiate between generic, relevant, and poor-performing keywords.

After optimizing the same account for 22 days, we increased overall relevant traffic by 15% and NTB orders by 30%. All of this was achieved after restructuring the campaigns and correcting the keyword and product targeting.

Just think how much potential the account had that was not fully explored due to AI-only optimizations.

For any account, I believe a combination of AI and human oversight is the best approach, as it can significantly enhance performance.