this is very much uncharted waters and openai is making changes to their models all the time

don't expect citation volatility to change anytime soon

seo has never guaranteed stability, and ai search will be even less forgiving

all you can do is stick to the basics: publish unique content that offers insights that aren't found anywhere else, be where the conversations are happening, and get ready to test, test, test

▶️ for your niche, track which domains surface month to month. learn what earns citations and replicate it
▶️ earn brand mentions on trusted and frequently cited sites where “trust already lives” - reddit, wikipedia, youtube, yelp, trustpilot, etc.

and if you’re trying to optimize or track ai citations, that’s a serious headache

our advice? play the long game:

▶️ build authority with insights llms can’t find elsewhere (original data, expert explainers, benchmarks)

3️⃣ model context, chat history, and model recency bias all influence what’s retrieved
4️⃣ openai regularly retrains or reweights retrieval data for freshness and diversity

because of these reasons, you might be cited in one conversation… and vanish in the next

unlike google search, where ranking signals are relatively stable, chatgpt’s retrieval pipeline is dynamic because:

1️⃣ it generates fan-out queries behind the scenes (and they change every time)
2️⃣ user prompts are often long, complex, and contextual, unlike most search queries

▶️ question queries? 57.9% trigger aio
▶️ non-branded is 1.9x more likely than branded
▶️ local and shopping are mostly safe (local 7.9%, shopping 3.2%)
▶️ but google is totally fine ai-ing YMYL: 44.1% of medical queries show an aio

study: ahrefs.com/blog/ai-over...

registration’s open now, but be quick, spots close nov 12 at midnight PST 👇
tinyurl.com/3w3wbep5

study: thecollegeinvestor.com/66208/37-of-...

throw in hallunications, and you have a recipe for disaster

remember: just because ai sounds confident, it's not a subject expert, and it certainly isn't a cerified financial planner

shocker, i know /s

would you trust a 63%-accurate financial advisor? probably not, right?

▶️ taxes
▶️ student loans / financial aid
▶️ insurance
▶️ product comparisons (it even made up products/rates)

and the cause is the same as ever: ai is summarizing the web

so, when the underlying content is outdated or flat out wrong, the ai answer are based on inaccuracies

their hot-off-the-press research found that 37% of aio's answers on personal finance are misleading, missing key info, or flat-out wrong

more than a one-third error rate in money topics is wild!

hello, ymyl?

here's where the college investor found aios struggle the most:

although, regularly updating your content could just earn you more citations

do with that what you will