Same conversation keeps coming up. Positions and impressions look basically unchanged, clicks are down, and then everyone argues about whether that is AI Overviews or whether the site got worse. Usually nobody can prove either. You can settle it with Search Console and a browser. No tool needed, takes about an hour. Pull your top 50 non brand queries by impressions. Search each one manually in an incognito window and note whether an AI Overview shows up today. Yes or no, one column. This is the tedious bit, budget forty minutes for it. Then pull CTR for those same 50 queries across two periods, whatever window you have before the drop and whatever you have after. Split them into two groups using your yes/no column and compare. If CTR fell in the AIO group and held in the other one, that is your answer. If it fell in both groups by about the same amount, AIO is not what happened and you need to look somewhere else entirely. And if it fell more in the non AIO group, you have a real SEO problem, and AIO was the excuse you were about to hand your client. Two things that trip people up when they run this. Average position stays flat while the actual placement moves. Position 3 with an AIO sitting above it is visually somewhere around where position 8 used to be, so the number in the dashboard can look fine while nobody sees you. And AIO presence is not stable. A query showing one today might not have shown one three months ago, so this tells you about right now rather than about the whole period. Getting the historical version properly needs a rank tracker that was logging AIO presence at the time, which most people were not running. I do the manual check rather than trusting a tool for it because the tools disagree with each other on whether an AIO fired, and when a client asks why traffic is down I would rather point at fifty screenshots than at a vendor's number I cannot explain. ...

A pattern Search Logistics has been looking at lately is sites holding roughly the same positions and impressions while organic clicks fall noticeably. Breaking the data down by brand vs non-brand, page intent, and query-level CTR helps, but it still gets messy when a couple of pages also lose rankings at the same time. What signals are you using to separate AI Overview-related click loss from an actual SEO decline?