Cameron Winslow.

Cameron's articles take one subject from different sides: what changes for a business when its customers put questions to AI engines instead of scrolling through search results, and what that does to the enquiries it receives. They sit in the AI Search section of the journal, and the list below adds each new one as it is published.

The article on five AI engines works from Brain Buddy AI's own monitoring records, collected between April and September 2026, and asks whether one blended visibility score can mean much when the engines disagree about which sources to trust. It treats engines that look things up as they answer, ChatGPT with search, Google's AI Overviews and Perplexity, as a different kind of result from Claude and Gemini answering from what they learned in training.

It then makes three narrower points: the engines cite different numbers of sources per answer, they favour different kinds of source, and being named in an answer is not the same as being linked from it, since a mention builds awareness while a link can send a visitor. It also names the risk in an averaged score: the figure can rise even as the enquiries flow to the businesses the other engines name.

The article on Google's AI Overview links looks at a change in Google's AI Overviews in September 2026, when links inside the answer text became far more common in one analytics company's tracking. Cameron reads it as a revenue question rather than a technical update, and puts a price on a lost enquiry using published Google Ads benchmarks for Australian family law firms and allied health clinics.

He compares Australia with the United States and the United Kingdom, where AI Overviews arrived earlier, and explains why the shift is hard to see in analytics: visits from ChatGPT can carry a tag that identifies them, while clicks from AI Overviews look like any other Google search visit. It closes with a short list for business leaders, which starts with answering customers' questions directly on your own site and includes making sure AI crawlers can reach it.

Both articles are built so a reader can check them:

• each ends with a list of its sources, most of them dated and linked
• the five-engines article labels its own monitoring data as such, and gives the dates and the sample behind it
• its limits get a section of their own, which says the businesses measured are clients being actively optimised, so their citation rates are no market baseline
• where a finding rests on one company's tracking, or the markets it covers are not stated, the AI Overviews article says that too
• unfinished work is marked as unfinished: the five-engines article mentions format experiments whose baselines are not complete, and publishes no numbers from them
• the same article warns that measuring AI Overviews from outside Google, Brain Buddy AI's own monitoring included, is likely to become less reliable, and explains why

One piece of advice appears in both: record where each enquiry came from at the point it arrives, whether by form, phone or booking, and ask every new customer where they heard about you. The five-engines article adds a second, to look at each engine on its own line rather than as one averaged figure.

To ask about something in one of his articles, use the contact page or email info@brainbuddy.ai, and name the article in your note.