How Perplexity, ChatGPT And Gemini Pick Their Sources

Why That Breaks the Old Playbook The old playbook assumed that if you occupied a high position, you got the visit. That link between position and visibility has weakened. Ahrefs looked at 15,000 long-tail prompts across four assistants in July 2025 and found roughly 80 percent of the cited pages did not rank for the original query at all.

Screenshots of favourable answers with no run count, which say nothing about how many attempts produced them. Impressions or traffic from unrelated channels included to fill a report. And activity described in the language of effort, such as ongoing optimisation, with no countable output attached.

Descriptions of your business moving from hedged to definite, which you can read yourself in the raw answers. Third party sources that previously described you wrongly now describing you correctly. And an increase in the fraction of runs naming you on buying intent prompts specifically, reported with run counts visible.

The Shared Architecture All three now commonly retrieve live sources rather than answering purely from training. Your question becomes one or more searches, a set of pages is fetched and read, and the answer is composed from what was read.

The first is accuracy. Somebody inside the business has to confirm that what gets published about your products, pricing and capabilities is true. The second is the third party work, which occasionally needs a decision only you can make, such as whether to engage with a critical review or approach a publication.

Why Local Is More Exposed The classic local query is a recommendation request with a geographic constraint, and that maps directly onto what a generated answer does well. Somebody asking who to call for a specific job in a specific town receives two or three names rather than a map and a list to work through.

If your category still gets meaningful traffic from those, a proposal scoped only to assistants will leave that work undone. Conversely, if somebody proposes an answer engine optimization programme and delivers only snippet optimisation, they are working on the older half of the definition.

A page worth having states what you do in that area specifically: which neighbourhoods, what travel time, what jobs are common there, what the local constraints are. If you cannot write anything genuinely local about a town, the honest answer is not to publish a page for it.

One check is worth running independently once a quarter, without telling anyone. Take ten prompts from the agreed set, run them yourself in a signed out session, and compare what you find against the most recent report. Broad agreement is reassuring. A consistent gap in the agency's favour is the single most informative finding available to you, and it is not something a report will ever surface.

How You Will Know It Is Working Ask for the raw answers, not a score. A credible report shows you the exact prompts, the exact text an assistant returned, and which pages were cited. You should be able to read it and form your own judgement without trusting anyone's index.

Because there is no independent scoreboard in this channel, an engagement can run for a year on the strength of a number the supplier produces. That is an unusual amount of trust to extend, and it makes knowing what to check more important here than in any other marketing channel.

The shortlist is shorter than a conventional local results page, which raises the stakes on being included. Being fourth on a map still gets calls. Being fourth in a recommendation that names three businesses gets none.

Some practitioners still use it that way, which makes it a superset of the newer work. Others use it as a synonym for the generative work specifically. Both usages are in circulation, which is why asking somebody what they mean by it is a reasonable question rather than a pedantic one.

Consistency Matters More Than Anywhere Else Local identity resolution depends on the business details agreeing across a long tail of directories, many of which nobody has looked at in years. Old addresses, disconnected numbers and previous trading names sit in these places indefinitely.

The Signals That Mean Something Four things are hard to fake and worth watching closely. Your own pages beginning to appear in cited sources, which is directly observable in any assistant that shows citations.

Generative Engine Optimization The broadest of the three in common use. It refers to being visible in systems that generate an answer rather than returning a list, which covers assistants, ai seo services summaries on results pages and any interface that synthesises rather than links.

The emphasis is on being included in a generated response, whether or not you are cited by name and whether or not it produces a click. The term appeared in academic work before agencies adopted it, which gives it slightly firmer footing than the alternatives.

How to Test Rather Than Trust Everything above is a starting hypothesis. Run twenty prompts in your own category across all three, from signed out sessions, recording the mode and the date, and count the cited domains for each.