Review currently available Google Ads recommendations and evaluate each one against the account's business goals, recent performance, conversion quality, campaign maturity, budget, bidding strategy, search intent, creative coverage, and change history. Treat optimization score uplift as context, not evidence that a recommendation should be applied. Be especially cautious with budget increases, looser targets, broad match expansion, network expansion, auto-created assets, and recommendations based on thin data. Group duplicates and ignore recommendations that are stale, incompatible with the campaign strategy, unsupported by account evidence, or impossible to judge safely because required business economics are missing. Submit only a small number of high-confidence, API-supported recommendation applications for administrator approval when their expected benefit and downside are clear. Summarize important recommendations that were rejected and why, but do not dismiss them unless the saved workflow explicitly asks for dismissal. If none are safe, make no changes.