The complaint arrives in nearly the same words every time: our rankings have not moved, our report looks the same as last year, and the pipeline is quieter. Nothing in a standard ranking report can explain that, because nothing in a standard ranking report looks above the rankings.
What changed
For a growing share of queries about IT services, accounting, insurance, benefits, financial advisory and similar categories, Google now composes an answer above the results and cites a handful of sources inside it. Some searchers read that answer and stop. Others read it, take the two or three names it surfaced, and search those companies directly. Either way, a position that used to sit near the top of the visible page now sits beneath a paragraph the buyer may never scroll past.
In parallel, a portion of buyers now begin with an assistant rather than a search engine. They describe the problem in their own words — not in keywords — and ask who they should call. The output is a short list. A company is on it or it is not.
What did not change
- Ranking still matters. Generated answers are assembled largely from sources that already perform in conventional search.
- Local still matters. Most professional-services prompts carry a place, and local data is what resolves them.
- Inquiry follow-up still decides everything. A named company that does not respond promptly loses the opportunity anyway.
- Directories still compete with companies, on this surface as on the last one.
The honest state of the evidence
We will not put a percentage on what a company has lost to AI Overviews. Analytics platforms do not cleanly separate a click that never happened, and the studies circulating vary widely by query type and industry. Anyone quoting a precise figure for professional services in New Jersey is extrapolating and not saying so.
What can be established for a specific company is narrower and more useful: which of its commercial queries currently return a generated answer, whether the company is cited within them, and who is cited instead. That is a finite audit, and it is where the conversation should start.
What to do about it
The work is not exotic. Consistent entity data across the site, the Business Profile and the directories. Pages that answer the question near the top rather than after three paragraphs of introduction. Named subject-matter expert authorship with real credentials attached. Presence on the third-party sources these systems disproportionately draw from. Every item on that list also improves conventional performance, which is the point — none of it is a bet on a single surface.
Then measure it: a defined prompt set, sampled on a schedule, reported as directional. That is a smaller claim than the market is currently making, and it is the one the evidence supports.