Uncategorized

Every platform reports share of voice and almost none of them define it the same way. Here is the definition worth using, the three choices that change the number, and how to measure it yourself without buying anything.

We already publish five methods for detecting brand mentions in AI answers, and that page says plainly what it does not cover: whether the things engines say about you are true. That is the half nobody monitors. A 5,000-prompt benchmark across five frontier models in April 2026 put factual recall hallucination between 4.2% and 12.7%, and citation hallucination between 6.8% and 19.1%. The IAB named factual inaccuracy rate as a metric brands should track. No platform we have checked reports it.

A SERP is the page you get back after a search. That definition has not changed, but what sits on the page has changed almost completely. The most useful way to understand one in 2026 is that it is not a ranking you hold. It is a page that re-renders, and three in four top-10 positions change URL daily.

Engines answer the same question differently each time, so every AI search tracking number is a sample with an error bar. Nobody in this category publishes that error bar. We worked it out from published tracking frequencies, and the result inverts the obvious assumption: moving up a tier buys you more prompts, not more confidence in any one of them. The number most teams act on weekly is the least reliable number the tool produces.

Independent researchers measured 21,143 citations to find what AI engines actually absorb into answers. Eighteen practices that follow from that evidence, plus four widely recommended ones with nothing behind them.

The short answer Every other overview of the top AI search engines is written for someone choosing a chatbot. This one is written for someone trying to be found inside one, which turns out to be a different question with a different answer. The engine that sends you the most traffic is not the engine […]

TL;DR: This is the most ambitious product we have ever launched. For three editions I have shown you where to find the questions you are losing, the sources AI trusts, and the repetitive work that quietly decides your results. Every one of them still ended the same way, with you going off to do something. […]

AEO proposals have converged on a standard deliverable list, and at least one item on it is something Google has publicly mythbusted. This guide covers what the work actually involves, what it costs at each tier, which line items to challenge, and how to run a procurement process that surfaces the difference.

TL;DR: Google holds YMYL content, Your Money or Your Life, to a higher trust bar in AI Overviews, and healthcare clears that bar differently than finance or legal content does. Only 24 percent of healthcare AI Overview citations overlap the organic top 10, meaning ranking well on Google and getting cited by its AI Overview […]