NetRanks

04/11/2026

Ask ChatGPT the same question twice and you’ll get two different answers. That’s why a screenshot of one answer tells you almost nothing about how AI talks about your brand.

Reha has answered the five questions marketing teams bring us most often, starting with why an assistant names a competitor instead of you and ending on whether it’s worth advertising inside ChatGPT. In between: how to measure representation when the answers keep moving, why correcting a wrong price on your own site doesn’t correct what the model says, and what you can honestly attribute to AI search right now.

Just under eight minutes, chaptered if you’d rather jump to one.

Ask ChatGPT the same question twice and you’ll get two different an...

08/10/2026

06/10/2026

AI search engines answer the same question differently every time, and they do not tell you why they named one brand and not another. NetRanks works out what to change so they name yours.

We treat this as an engineering problem rather than a marketing one. Our system measures how the engines behave across hundreds of thousands of answers, then traces each mention back to what produced it, down to the individual sentence an engine drew on (whether taken from your own site or from third-party pages you don’t control).

Whether an engine names you comes out of the trust signals in those sources. The Answer Impact Model works out which of those signals are broken and which are worth strengthening, so that what comes back is a prioritised roadmap rather than a list of findings.

You can also test before you publish. Stage a change inside the platform and see how ChatGPT or Gemini would read it, instead of shipping it and waiting a month to find out. And you can track up to 250 competitors across engines, topics and regions, so you can see who is being named in your place.

Follow NetRanks for our research on how AI search engines work and what brands have to do to earn their trust.

02/10/2026

Ask ChatGPT the same question twice and you will likely get two different answers. Ask Google the same question twice and you get ninety-nine percent the same answer.

A traditional search engine (Google) is, in effect, a lookup: the same input yields the same ranked output, and an optimizer can reverse-engineer the rules and climb. A language model (ChatGPT, Gemini, Perplexity, Claude and the like) answers probabilistically, drawing on a shifting internal sense of what is relevant and credible. And it also carries context: language, region, industry, as well as your own context assembled during your conversations with the engine.

So a screenshot of “what ChatGPT says about my brand” is, in a sense, meaningless. The only way to know how an engine behaves is to ask at scale and read the distribution.

29/09/2026

Do ChatGPT, Gemini and Perplexity cite the same websites? Between 13 August and 13 September we tracked 9 question sets across ChatGPT, Gemini and Perplexity: 302,920 answers and 3,491,564 source links. For this study, 87% of the answers came from a B2B software question sets (based on other data we have, the platform mix varies a lot by industry).

It is clear that each engine has its own preference when it comes to sourcing the answers. This is why a single AI visibility score is worth so little. If you are deciding where to publish, the first thing to decide on is which engine you care about.

NetRanks shapes how ChatGPT, Gemini and Perplexity talk about your brand, engine by engine.

24/09/2026

Every company has someone responsible for what the press says about it, and someone responsible for how it is indexed on Google. Almost nobody is responsible for what ChatGPT and other AI engines say.

When a buyer asks ChatGPT or Perplexity to recommend a supplier, the model builds a shortlist based on the trust signals it picks up on. AI relations is the job of building trust with AI engines to make sure your company is mentioned where, how, and how often you’d like it to be mentioned. Similar to public relations, but pointed at a machine, instead of a journalist.

At NetRanks, we think this will be a role on the org chart within the next couple of years, probably sooner. Link in bio to learn where your company currently stands.

22/09/2026

NetRanks analyzed 945,150 answers from ChatGPT, Gemini and Perplexity, and the 10.3 million source links behind them, between 15 June and 13 September 2026, to see how often each engine cited YouTube.

On Gemini, YouTube is the most cited website of all. On ChatGPT it ranks 1,813th, and ChatGPT didn’t cite a single YouTube page until 30 July. Perplexity cited YouTube heavily until 12 August, when its YouTube links fell from 1,711 in a day to 221.

The chart shows the 9 question sets tracked on all three engines for all 13 weeks, 853,320 answers in total.

NetRanks shapes how ChatGPT, Gemini and Perplexity talk about your brand, engine by engine.

21/09/2026

Is YouTube still a source in AI answers? NetRanks read 945,150 answers and analyzed 10.3 million source links from ChatGPT, Gemini and Perplexity between 15 June and 13 September 2026, and the answer depends on which engine you ask.

- On Gemini, YouTube is the most cited website of all, first among the 24,838 websites Gemini cited repeatedly. About 27% of Gemini answers included a YouTube video in June, and the same share still did in September.

- On Perplexity, YouTube citations dropped overnight. On 11 August it returned 1,711 YouTube links, and on 12 August 221, even though its total number of links went up. Instagram and Facebook fell almost as sharply, while Reddit, LinkedIn, Wikipedia and Medium held steady or grew.

- ChatGPT did not cite a single YouTube page among more than half a million sources until 30 July. It does now, in about one answer in 2,000.

When an AI engine does cite YouTube, it cites one video. 98% of YouTube citations point to a single watch page, and channel pages get under half a percent.

🔍 We are conducting more experiments in the weeks to come. What would you like to get insight into?

NetRanks shapes how ChatGPT, Gemini and Perplexity talk about brands, engine by engine. Link in bio to find out where your brand stands.

17/09/2026

An early customer of ours has hundreds of millions in revenue and a permanent place on Google's first page; but across a thousand questions put to the AI engines, it surfaced two or three times.

NetRanks collected the data, tied it to a single product, and returned about a hundred and thirty precise changes that would make them trustworthy to ChatGPT and the likes. After the changes had been implemented, two or three mentions per thousand had become four hundred, making the brand one of the three most frequently named in its category.

Position lagged behind frequency: at first the model named it only at the foot of its answers, an afterthought, and it took a further month of iteration before the brand appeared among the first names offered. Visibility rose first, then traffic, then leads.

Trust is accrued, not switched on.

15/09/2026

Is Your Own Website Breaking AI's Trust?

When an AI assembles a recommendation, it typically reads a set of sources and often discloses which ones it leaned on. For one company, the single most-used source was the company's own website: the machine had found them, read them, and treated their site as the authority on the question. Then it recommended their competitors.

Their content was missing the right trust signals. So every time AI read their website, it broke its own trust and recommended a competitor. They had the power to control the answer and were using it to recommend other brands. Their own website was sabotaging their visibility.

The site had been built, like almost every corporate website, for two readers: people, who want something legible and attractive, and traditional search engines, which want something indexable. It had never been written for an AI that reads in order to decide whom to trust. The logical conclusion is almost absurd: this company would have been more visible with no website at all.

There is no agreed name for the phenomenon yet (at NetRanks we call it "negative signalling," or "trust-breaking content") and it cannot be seen without the right instruments. A company cannot simply read its own pages and spot the fault, because the fault exists only in how a model interprets them.

https://youtu.be/YdVoBB3vzJg

When an AI assembles a recommendation, it typically reads a set of ...