Short answer, as of September 2026: a new site can earn AI citations without waiting years for domain authority, because assistants select passages that answer a question and can be corroborated, not domains that rank well. The binding constraint on a new site is not age. It is that nothing outside your own website says you exist, and that is a problem you can work on directly.
This is one of the few places where AI search is genuinely kinder than traditional search. Competitive rankings really do reward accumulated link equity, and a new domain really is disadvantaged for years. Citation works differently enough that the sequencing advice changes.
Why a new domain is not automatically excluded
When an assistant answers a question, it retrieves candidate passages and decides which to lean on. The decision weighs whether the passage directly answers the question, whether it is specific, whether the source is identifiable, and whether other sources agree. Domain age is not a step in that process — it is correlated with the things that are, which is a different claim. The selection mechanics are set out in how ChatGPT, Perplexity, Gemini and Copilot pick sources.
That is why a six-month-old site with one genuinely specific page on a narrow question can be named in an answer while a large established competitor with vague marketing copy is not. Specificity is available to you immediately. Authority is not.
The honest counterweight: corroboration does a lot of work in that process, and a brand-new business has none. Nobody has reviewed you, no directory lists you, no forum mentions you. That is the real gap, and it is the one to attack first.
The order that works
The sequence matters more than the individual tactics, because several of these unblock the others.
Weeks 1-2 — make yourself retrievable. Nothing else counts if assistants cannot fetch your pages. Confirm the AI crawler user-agents are allowed in robots.txt and, more often the culprit, not silently blocked by your WAF or bot protection — a check most new sites fail without knowing, covered step by step in AI crawler access. Then confirm your content is present in server-rendered HTML rather than assembled client-side, since a page that needs JavaScript to show its text is a page some retrievers see as empty; the detail is in JavaScript rendering and AI visibility.
Weeks 2-3 — make yourself identifiable. Publish the boring facts in machine-readable form: who you are, what you do, where, what it costs, how to contact you. Organisation and service structured data, a complete about page, consistent naming. New businesses routinely skip this because it feels like paperwork rather than marketing. It is the paperwork that lets an engine name you confidently instead of describing you vaguely. See structured data and llms.txt.
Weeks 3-8 — publish the specific things nobody else will. This is where a small site wins. Established competitors write broad, hedged, committee-approved content. You can publish exact prices, exact process, exact limitations, exact timelines. Specificity is the currency, and the argument for publishing numbers — including the commercial objections — is in publishing prices and cost transparency.
Weeks 4-12 — build corroboration. In parallel, not afterwards. Get listed in the directories that matter for your category, get your profile complete and verified, and start earning reviews from real customers. This is the slow part and the reason to start it early; third-party sources and AI citations covers what actually moves.
Week 8 onward — measure, then iterate. Baseline your prompt set before you conclude anything about progress, and expect the timeline to run in months rather than weeks; the realistic curve is in how long GEO takes.
What a new site can and cannot compress
| Factor | Can a new site shortcut it? | Realistic timeline | Notes |
|---|---|---|---|
| Crawler access | Yes, immediately | Same day | Config change, not a growth problem |
| Server-rendered content | Yes, immediately | Days to weeks | Depends on your stack, not your age |
| Structured data and entity facts | Yes, immediately | Days | Nearly always the fastest win available |
| Passage-level specificity | Yes | Weeks | Limited by willingness, not authority |
| Indexing and discovery | Partly | Days to weeks | Notifying on change helps; being crawled at all is the gate |
| Reviews and third-party mentions | No | Months | Requires real customers and real time |
| Broad topical reputation | No | Many months | The genuine authority constraint |
The pattern in that table is the strategy. Everything a new site can compress, it should compress in the first month, because those items cost configuration and honesty rather than time. Everything it cannot compress should be started in the first month too, precisely because it is slow.
Being crawled at all deserves its own note. New sites are discovered slowly, and notifying search infrastructure when pages change is cheap and helps; the practical mechanics are in IndexNow and indexing speed.
Pick a narrow enough question to actually win
The most common strategic mistake a new site makes is competing on the broad head question in its category on day one. "Best CRM software" is not winnable, and the attempt produces a page too generic to be quoted.
The alternative is to find questions where your specific answer is better than anyone's general answer. Those cluster in three places: narrow qualifiers (a specific integration, constraint, industry or edge case), operational specifics (what it costs, how long it takes, what happens if something goes wrong) and honest comparisons including where you are the wrong choice.
That last one is uncomfortable and effective. An assistant asked "is X right for a two-person team" is looking for a source that draws a boundary. A page that says plainly who the product does not suit is far more likely to be quoted than one claiming to suit everyone — and it converts better, because the people it turns away were never going to stay.
Choosing the questions is itself the first real deliverable; the sourcing method is in how to build a prompt set worth tracking.
Three shortcuts that waste the first six months
Publishing volume instead of specificity. Forty thin posts on broad topics from a new domain produces nothing. Five pages that answer narrow questions with real numbers can produce citations. Volume is the instinct because it feels like progress; it is the slowest possible route.
Buying links and expecting citations. Link-buying targets a ranking signal, not a retrieval decision, and it carries the same risk it always did. Corroboration comes from being genuinely present where your category is discussed, not from purchased placements.
Declaring failure at week six. Citation lags publication, because engines have to crawl, index and then be asked. Six weeks of no movement on a new domain is expected, not evidence the approach is wrong. Decide your review point in advance — a quarter is reasonable — and hold to it.
A realistic first quarter
If you do only the sequence above, a plausible outcome after three months on a new domain is: crawler access confirmed and monitored, entity facts published and consistent, five to ten genuinely specific pages live, a complete verified Business Profile, a handful of real reviews, a baselined prompt set with dated answers, and the first citations appearing on your narrowest questions.
That is not a spectacular-sounding quarter. It is the one that compounds, because every item on it is a prerequisite for the next one being worth doing.
You can baseline and track that progress on the Starter plan, which covers one brand and 200 prompt-checks a month — enough for a small set run repeatedly, which is what a new site needs. The features overview shows what the tracker records, and the honest first conversation to have about expectations is the one in when a client first asks about AI search.
Frequently asked questions
Can a new website get cited by AI without domain authority? Yes. Assistants select passages that directly answer a question and can be corroborated, and domain age is not a step in that decision. A new site with one genuinely specific page can be cited where a large competitor with vague copy is not. The real constraint is corroboration, not age.
How long before a new site sees its first AI citation? Expect months rather than weeks, because pages must be crawled and indexed before an engine can be asked and lean on them. Six weeks with no movement on a new domain is normal. Set a review point a quarter out and judge it then.
What should a brand-new site do first for AI search? Confirm AI crawlers are not blocked by robots.txt or your WAF, and confirm your content is in server-rendered HTML. Both are configuration changes that take hours, and nothing else you publish matters if retrievers cannot fetch and read the page.
Do I need backlinks to be cited by AI assistants? Not in the way you need them to rank competitively. What matters is corroboration, which comes from directories, reviews, profiles and genuine mentions where your category is discussed. Buying links targets a ranking signal rather than a retrieval decision and carries the same risks it always did.
Should a new site target broad or narrow questions? Narrow. Broad head questions are dominated by established sources and produce pages too generic to quote. Target narrow qualifiers, operational specifics like cost and timing, and honest comparisons that state who you are not right for.
Is publishing more content faster the right strategy for a new domain? No. Forty thin posts on broad topics produce nothing, while five pages answering narrow questions with real numbers can produce citations. Specificity beats volume, particularly for a site with no accumulated reputation to lend weight to generic pages.
What can a new site never shortcut? Reviews, third-party mentions and broad topical reputation. All three require real customers and real elapsed time. Start them in month one precisely because they are slow, and compress everything else immediately.
Does publishing prices really help a new business get cited? It is one of the most effective moves available, because cost questions are extremely common and most competitors refuse to answer them on the page. It is also a genuine commercial tradeoff worth deciding deliberately rather than by default.