One answer in sixteen: how often AI search names a young company

In the first wave of our cohort study, young YC companies were named in 6.4% of usable AI search answers, roughly one answer in sixteen. The established comparison group was named in 36.7%. The question construction and group definitions matter to that gap.

What we measured

A cohort of 118 Y Combinator companies from the 2021 to 2026 batches, drawn by a seeded random draw from YC's public directory, and 18 of YC's own designated top companies as the control. Both groups come from the same institution. That narrows the comparison, but does not hold company size, funding or founder quality constant. The full study also included 16 separately chosen companies in our category; those companies are excluded from the young-versus-graduated comparison below.

Each company got thirty customer questions built from its own homepage with the brand removed. Google AI Overviews and Perplexity received all thirty; Gemini and ChatGPT each received the same eight-question subset. That is 30 + 30 + 8 + 8 = 76 planned answers per company. A successful Google search with no AI Overview stays in the denominator as a result that did not name the company. Collection ran across 22 and 23 August, including re-runs after errors. A mention is the company's name appearing in the answer. Detection is recall-first: a close call counts as a mention, which can overcount mentions in either group.

The number

The pooled rates differ by 30.3 percentage points, or about 5.7 times. Every engine showed a gap: young-company mention rates ranged from 6.0% to 6.8%, and the established group from 30.6% to 39.4%.

The distribution says more than the average. Forty-eight of the 118 young companies were never named in the usable answers about what they sell. Young companies had 67 to 76 usable answers each; failed requests were excluded, not counted as silence. One young company in 118 cleared 50%. Seven of the eighteen graduates did.

For the shared eight-question subset, we compared only company-question pairs with a usable result from every engine. None named the company in 756 of 900 young-company pairs (84%), versus 57 of 143 graduate pairs (40%). Pairs with an engine error were excluded from this comparison.

Tap to zoom

How batch years compare

The young group spans six batch years, with 18 to 21 companies in each. This is a comparison of different companies at one point in time, not a record of how the same companies changed as they aged.

The 2025 and 2026 batches average around 2% to 3%, and roughly two thirds of the newest batch was never named. The 2022 and 2023 batches sit near one answer in ten; the 2021 batch is lower. That pattern does not show that waiting improves visibility, or that improvement stops at a particular age. The groups differ in survival, size and other characteristics that this comparison does not isolate.

Tap to zoom

Who clears the bar young

Nine young companies were named in at least a quarter of their answers. Reviewing that small group suggested category specificity and public technical material as possible follow-up questions. The study did not isolate either factor. One young company, a 2022 fintech API, exceeded 50%.

The aggregate mention rates for young companies are close across engines. That does not mean the engines name the same companies or agree on individual questions. The pooled figures also give more weight to Google AI Overviews and Perplexity, which received more questions; only the shared eight-question subset permits a comparison using identical questions.

What this does and does not say

• It is YC. It does not generalise to startups outside YC, and no figure here claims it does.
• The frame contains only survivors: YC marks the rest acquired or inactive. The older the batch, the more selected it is. Roughly three in ten of the 2021 batch is already out of the comparison, so the 2021 figure is about the companies that made it.
• The question is the instrument. Every figure is from questions built from homepage text with the brand removed. A different construction gives different numbers; our own pilot showed the swing between two constructions can be sevenfold. The construction is versioned.
• Perplexity's answers for most companies were collected some hours after the other engines', after an account quota lapse mid-wave was topped up and re-fired. Every pooled figure excludes errored answers rather than counting them as zero.
• The control leg is 18 companies. Its per-stratum figures are directional.

The questions to ask of any figure like this

Where did the questions come from? How many answers, not scans? What counts as a mention, and how generous is it? Who was in the sample, and how old are they?

AfterLaunch helps founders track where customers can find their company and turn those findings into growth work to review and complete.