What is AI citation readiness, and what does the score measure?

AI citation readiness is how likely a page is to be retrieved and quoted when an answer engine assembles a reply to a question that page should own. The score measures the page rather than the answer. A citation count tells you what happened last week; a readiness score tells you which properties of the page decided it, which is the only one of the two you can act on directly.

The published readiness scores agree on the shape of the number. They run 0 to 100, they are weighted averages of category sub-scores covering structure, authority signals, factual specificity and crawlability, and they come with reading bands: below 40 reads as major blocking problems where an engine either cannot reach the site or finds nothing to trust, 70 to 84 as above average and cited with reasonable frequency, and 85 and over as consistent citation across platforms, according to SearchScore’s published guide to the AI visibility readiness score, read 2 October 2026. SIGNALS reports on the same 0 to 100 scale for that reason.

What the frameworks do not agree on is what goes inside the average, and that is the distinction worth knowing before comparing two scores. One published formula weights entity resolution rate at 0.25, AI mention rate at 0.20, citation rate at 0.20, source-role evidence coverage at 0.20 and cross-engine consistency at 0.15, according to the same guide. Three of those five are measured in the engines’ answers rather than on the page, so a score built that way moves when the engines rotate their sources even if nothing about the site changed. The seven SIGNALS dimensions are all read off the page, and the citation outcomes are kept in a separate series that is re-run on a schedule. That is a difference in what is being counted rather than a claim about which number is better: a score built from outcomes tells you where you stand this week, and a score built from the page tells you what to change.

The research behind the score

Each paper below contributed specific findings that shaped the SIGNALS framework, the weighted search signal framework this practice scores against. The weights aren't assigned subjectively — they're derived from measured effect sizes in peer-reviewed studies. When the research says something carries a 35% weight, it's because the measured effect size is 35% stronger than the next strongest signal, not because we thought it sounded right.

The 7 SIGNALS dimensions

Each dimension is weighted by its measured effect on AI citation frequency. The weights sum to 100%. Alignment carries 35% because it's the only signal with a documented causal effect that survives domain-level controls, according to Discovered Labs' citation analysis. Not because we think content is more important than structure, but because the research is unambiguous about it.

	Dimension	Weight	What it measures	Source
S	Structure	12%	Heading hierarchy (H1→H2→H3), TL;DR boxes, self-contained sections, comparison tables. 68.7% of cited pages use logical heading structure.	ConvertMate 2026 AgentGEO 2026
I	Intent	10%	Coverage of adjacent buyer intents and query fan-out. A single user prompt decomposes into multiple sub-queries — a page must address several to be cited across all decompositions.	Princeton GEO 2024
G	Grounding	13%	Verifiable claims: statistics with named sources, inline citations, expert quotes. Princeton measured +41% citation frequency from sourced statistics, +28% from named expert quotes.	Princeton GEO 2024
N	Newness	5%	Visible Last Updated timestamps, specific year references, internal date consistency. Page age effect is real but small: β=+0.05 — meaningful but not a priority fix unless everything else is already addressed.	Discovered Labs 2026 β=+0.05
A	Alignment Dominant signal	35%	Does the vocabulary on this page mirror how buyers actually search? Measures lexical overlap between page content and documented buyer search patterns. The only signal with a documented causal effect independent of domain authority.	Discovered Labs 2026 β=+0.37, q≈10⁻⁷³
L	Language	10%	Buyer-phrased title and H2s, fluency score, title-to-prompt similarity. Fluency optimization increases AI visibility 15–30%. Title-prompt similarity has a measured effect of β=+0.09.	Princeton GEO 2024 β=+0.09
S	Substantiation	15%	Third-party validation: G2/Capterra ratings, Reddit mentions, press coverage, named author credentials. Brands mentioned on third-party domains receive 6.5× more AI citations than brands existing only on their own site.	ConvertMate 2026 Discovered Labs 2026

How does SIGNALS measure AI citation decay?

SIGNALS measures AI citation decay as the difference between two runs of the same prompt set, never from a single check. Each run records the prompt, the engine, the date and every URL the answer cited, which turns a page's citation history into a series you can plot. Decay is the downward slope in that series. Two numbers come off it: citation half-life, the time a page takes to fall to half its peak citation frequency, and the one-and-done rate, the share of cited URLs that appear in one run and never again.

Re-measurement belongs in the methodology rather than beside it, because the published decay rates are fast. Trakkr Research logged 108,650 citation URLs across 10,991 brands and eight tracked models, running identical prompts daily, and found that 73.5% of citation URLs appeared exactly once and did not return, with the median brand falling to half its peak citation count in 31 days, in its write-up "The Half-Life of AI Citations", checked 21 September 2026. A score measured once describes a page for about a month.

How much movement counts as decay depends on how often a page is cited at all, so the reading starts by sizing the noise. BrightEdge's citation volatility analysis found that frequently cited domains moved about 0.7% week to week while sporadically cited domains swung by more than 50%, a 70-fold gap that held across ChatGPT, Perplexity, Google AI Overviews and AI Mode, and it put the stability threshold near 50 citations, where weekly volatility drops from roughly 50% to 8%, according to BrightEdge, "AI Search Engine Citation Volatility Analysis", checked 21 September 2026. Below that threshold a missing citation is one sample rather than a trend, and a rewrite ordered on that evidence is work nobody needed.

What is each re-measurement run for?

Each run in the cadence answers a different question, and running the wrong one produces a number nobody can act on. The four below are what a SIGNALS engagement actually schedules.

Run	When it happens	What it answers	What counts as a real move
Baseline	Twice, a week apart, before any fix	Which engines cite you now, on which prompts, and how far two identical runs differ	Nothing yet. The second run is there to measure your own noise
Verification	As soon as the fixes are live	Whether the structural change registered at all	Appearing in the retrieved set, not the position inside the answer
Tracking	Monthly through an engagement	Whether citation frequency is rising or falling	The same direction across three consecutive runs
Maintenance	Quarterly once the work stops	Whether a page that won has started to slide	A shortening half-life, or a one-and-done rate that climbs

Monthly tracking follows from the two figures above rather than from a reporting habit. Structural changes take two to six weeks to show up in citation frequency, so a run a fortnight after a fix can only tell you whether the page is being retrieved again, not whether it is winning. A median half-life of about a month means a quarterly cadence meets a sliding page after it has already lost most of what it had. Weekly runs earn their cost on prompts with a buyer behind them, and only there.

The arithmetic, worked through

Take a page cited in 8 of 40 prompt runs in the month it peaked, a citation frequency of 20%. Three months later the same 40 prompts cite it 4 times, or 10%. Half the peak in three months is a three-month half-life, and on that curve the page is cited in 5% of runs three months after that, which is one prompt in twenty. The slope is the number to act on rather than the 10%, because the slope says how long the page has before it stops appearing in answers it used to own. Run the same arithmetic on a page cited twice at its peak and it tells you nothing, which is what the volatility threshold above is for. These figures are arithmetic rather than a measurement of any page, and the metric definitions behind them are set out in our guide to measuring AI citation decay.

A decay curve does not say why a page is sliding, and three different failures draw the same line. The page may have dropped out of retrieval. It may still be retrieved and no longer chosen. Or the question moved, and the page now answers a phrasing nobody types. Telling them apart means reading what replaced you in the answer, which stays a manual step in this methodology. Tracking AI visibility across the four engines covers the prompt-set design underneath all of it.

Which AI citation metrics sit outside the readiness score?

The AI citation metrics that measure outcomes are kept outside the readiness score and reported as a separate series, because mixing them produces a number that cannot be read. Four of them matter in an engagement, and each one answers a question the readiness score cannot.

Metric	Measured where	What it answers
Citation rate	In the answers, per round	What share of the prompt set names you at all
Share of voice	In the answers, per round	How you stand against the firms named alongside you
Citation half-life	Across rounds	How fast a win erodes once it is earned
One-and-done rate	Across rounds	Whether a cited page holds its place or appeared once
Readiness score	On the page	Which properties of the page to change next

Keeping the two apart is what makes either readable. Fold a citation rate into a readiness score and a week when the engines rotated their sources looks identical to a week when the page got worse, which is the failure the separate series exists to prevent. The arithmetic behind the last three rows, including how to compute a half-life from a log and how to size your own noise before reading a trend, is worked through in how to measure AI citation decay, and the prompt-set design that produces the denominator is in how to track AI visibility across engines.

How SIGNALS compares to other tools

Standard SEO tools were built for Google's PageRank. AI visibility trackers measure citation outcomes — they tell you whether you're being cited, not why or what to fix. SIGNALS is built to diagnose and fix the structural causes, not just report the symptom.

Read the three columns as three jobs rather than as three products competing for the same slot. A rank tracker answers where you stand in classic search, an outcome tracker answers whether the engines are naming you this month, and a page-level diagnostic answers which property of which page to change next. Most programmes end up holding two of the three, because a diagnosis with no outcome series cannot tell you whether a fix landed, and an outcome series with no diagnosis leaves you watching a number you have no lever on. Which of the tracking tools fits which situation is compared separately in the 2026 AEO tools comparison.

Capability	SEO tools (Ahrefs, Semrush)	AI trackers (Profound, Trackr)	SIGNALS
Measures AI citation readiness	✗	✗ (outcomes, not causes)	✓
Diagnoses why AI ignores a page	✗	✗	✓ Pipeline diagnosis
Generates page-specific fixes	✗	✗	✓ BLUF, FAQ, H2s, schema
Implements the fixes on your pages	✗	✗	✓ WordPress, Shopify, Webflow
Based on peer-reviewed research	Partially	Partially	✓ 4 papers, 2M+ citations
Works for small sites	✗ (favors high-DA domains)	✗ (only tracks top brands)	✓ Page-level signals, no DA bias
How you pay	Subscription only	Subscription only	Done-for-you engagement, scoped from a free visibility assessment

Frequently asked questions