Optimizing your website gets you found. It does not get you recommended.
Every month we ask AI answer engines the same twelve questions about our own product, five times each, with web search on, and publish what they say. Two of the questions use our name. Ten are the questions a buyer asks when they want a tool and have never heard of us. Four months in, those two halves tell two different stories, and the gap between them is the most useful thing we have measured.
Found by name: this is the part a website controls
Asked about AEO Analyzers by name, the engines returned our site in 55% of answers in August, 80% in September, 88% in October and 78% in November, 40 answers each month. (August is the restated figure; the ledger explains why.) In November our own site was among the sources cited in 31 of those 40 answers. Claude, Perplexity and Gemini returned it every time. ChatGPT returned it 1 time in 10, against 5 of 10 in October.
Being reachable, being readable and saying plainly what the product is: that is site work, and this is the column where it shows. It is not a straight line. It went up three months running and down in the fourth.
Recommended to buyers: zero, every month
On the ten category questions, the engines recommended us in 0 of 200 answers in August, 0 of 200 in September, 0 of 200 in October and 0 of 200 in November. That is 0 of 800 across four months. The word "aeoanalyzers" appears in none of those answers. We are not ranked low on those questions. We are not in them.
Where a recommendation comes from
To answer a category question, an engine searches, reads a handful of pages and builds its answer from what those pages say. So in November we fetched the 40 pages the engines read most often for our category questions and checked each one for our name. None of the 40 names us. Four could not be read.
That pattern is not ours alone. In our benchmark of 11 other companies, using the latest four-engine sweep of each one (July 26 to October 8, 2026), the median company was found by name in 100% of answers and recommended in 28% of category answers. Companies are not named, and our own products are not included.
What the ledger says, read plainly
Site work is how a company gets found and described. It is necessary, and in our own numbers it is visible. A recommendation is assembled from the pages the engines read for the category question. Those pages are not ours, and today none of them mentions us. Nothing on our own site can put our name on someone else's page.
So the work we are doing now is on those pages, and on the places the engines read when they look a company up. From our Record, with dates:
- October 15th, 2026: Wikidata entries created for the company and the product; AlternativeTo listing live.
- November 1st, 2026: Crunchbase profile live; LinkedIn company page live; SourceForge and Slashdot listing submitted; Crunchbase and AlternativeTo added to the Wikidata entries.
What we will watch in December
The December measurement uses the same twelve questions, the same engines and the same five runs. It will report found by name, described accurately, recommended to buyers and Source coverage, next to the dates above. If the recommended column moves off zero, the ledger will show the answers it came from. If it does not move, the ledger will show that too. Dates show timing, not cause.
Every month so far, with the stored answers behind each figure, is on the ledger.
AEO Analyzers (aeoanalyzers.com) was created and is solely maintained by Lindsay Hiebert, founder of PIGENAI LLC. It is unaffiliated with any similarly named browser extension, plugin, or tool.