Ranking in AI search in 2026 means being present, readable and consistent across the open web, because the model draws on what it already knows about your entity before it ever searches and then decides which sources it will quote. That is the whole game in one sentence, and it is why a portfolio of aged domains with clean histories can outperform a brand new site with better design. The SEO.Domains Mastery Summit in Sofia exists precisely because this shift is now an operational problem rather than a theoretical one: around 300 SEOs, affiliates and agency owners gather for a mastermind day on 9 September, then two days of main-stage sessions, and the agenda covers aged domains, PBNs, authority transfer and LLM visibility. Everyone there is working the same seam. Very few are working it in the right order.
A-S-S stands for Authority, Sources, Specificity, and it maps directly onto the three stages a generative engine moves through when someone asks it a question. Before it carries out a search, the model already knows certain things about you, which is what authority refers to. Sources are the result of its search. Specificity is how precisely your page answers the exact question that was asked. Nothing about this is mystical, and nothing about it is new to anyone who has built links for a living, but the weighting has changed. If you want the longer treatment, the full A-S-S walkthrough (https://www.youtube.com/watch?v=FZu4NB-2EhA) covers the framework end to end. If you want to pressure-test your own portfolio against it, a ClickBomb strategy call (https://seojesus.com/clickbomb-strategy-call/) is the place to bring your specific domains. For the operator community around it, there is also the Church of SEO Jesus (https://www.skool.com/church-of-seo-jesus).
The uncomfortable part for domain investors is that A-S-S does not respond to the levers most of us grew up pulling. Exact-match anchors, domain age alone, and a clever template do not move Authority. They might move Sources, but Sources are only the middle third.
Work through these in order. Skipping steps is how portfolios with genuinely good domains end up invisible.
Source selection in generative engines is volatile in a way that classic rankings are not. A repeat of the same question on the same day resulted in a change to 60.4% of the sources the model used. Read that again as an operator: the same prompt, the same day, and a majority of cited sources shifted. Any strategy that assumes you can lock a citation slot the way you once locked a position three is built on sand.
This is the part the summit circuit tends to underplay, because it is easier to sell a method than a volatility profile. What it means practically is that you should be optimising for presence across a wide set of plausible source surfaces rather than for ownership of any one of them. Generative Engine Optimization optimises for the answers AI search engines give, not only the ten blue links, and that reframing matters most when the underlying answer set is a moving target. You are not defending a rank. You are trying to be in enough of the candidate pools often enough that the reshuffle favours you more than it hurts you.
There is a second reason volatility cuts against lazy portfolios. A single authority hub that gets picked up on one day may vanish on the next, so the domains that keep resurfacing are the ones with genuine, independently verifiable entity footprints. That is the quiet argument for aged domains with clean histories: the entity record has had years to accumulate the consistent cross-web mentions verification depends on, and that layer does not reshuffle when the retrieval does.
| Layer | What it depends on | Volatility |
|---|---|---|
| Authority | Consistent cross-web mentions and entity records | Slow to change, durable |
| Sources | Which pages the engine retrieves and cites | High, 60.4% in one same-day test |
| Specificity | Readable, question-matched answer blocks | Low, fully under your control |
The two layers you can control outright are Authority and Specificity, so that is where the checklist time should go, and the volatile middle layer is where you measure rather than manage.
Visibility tools are useful for directional signals and useless as evidence. The 1.3% to 1.8% agreement figure between API-based monitors and a recorded session is not a knock on those products so much as a boundary condition: they are querying a different surface with a different context window and calling the output a measurement. Domain investors in particular tend to over-buy dashboards because portfolios are large and dashboards scale better than patience.
If you take one operational habit from the summit's themes, take this one: once a month, sit down and manually run your twenty most commercially important questions, record the sessions, and note which domains are cited. That log will be more useful to you than any SaaS export, and it will be the raw material for every content and acquisition decision you make afterwards.
Expect the Specificity layer to respond within weeks, because readable, question-matched answer blocks get picked up relatively quickly once crawlers reprocess the page. The Authority layer moves on the timescale of entity consolidation, which is months, and it is the layer that determines whether you keep reappearing after a reshuffle. The Sources layer is not something you wait for at all, since it can change within a single day and cannot be held.
Links help, but entity records and consistent cross-web mentions carry a share of the load that would surprise anyone raised on anchor text as the primary currency. With 37% of recommendations in one study coming from directories and review sites, and verification resting on consistent mentions across the web, treating link acquisition as the only lever leaves a large part of the opportunity unworked. On a domain portfolio, the more relevant question is whether each asset has a coherent, verifiable identity at all.
An aged domain with a clean history starts with a real advantage, because the entity record already exists and the cross-web mentions are already consistent, which is exactly what verification rewards. That advantage is forfeited the moment you point the domain at thin, JavaScript-dependent content, because you have solved the Authority problem while reintroducing the Specificity problem. Age is a head start, not an exemption. The summit's agenda on aged domains, PBNs and authority transfer exists because this is where most portfolio value is either realised or wasted.
Pick your five most commercially important domains and run the render test before anything else. If a model cannot read the answer without executing your code, stop and fix that, because every other item on the checklist assumes a readable page. Then audit each asset's entity record against the directory and review surfaces that fed 37% of the recommendations in that 82-answer study, correcting name, address and description until they match exactly across every listing you control. Then rewrite the headings as questions and put the answer in the first two sentences underneath.
Only once those three steps are done should you start measuring sources, and when you do, measure by recording sessions manually rather than reading a dashboard, given how poorly the monitors tracked. The floor is the thing. AI search does not reward the loudest portfolio, it rewards the most legible one, and legibility is assembled in a boring, sequential order that most operators skip because the middle step, source chasing, feels more like progress. Work the baseline, hold it consistently, and the reshuffles become a source of upside rather than a source of anxiety.
