We Built a Research Tool for Journalists Who Are Tired of Chasing Our Citations
Every few weeks, a reporter working a story about AI search asks us some version of the same question: where does that number come from?
It's a fair question. Our corpus is large — dozens of working papers, hundreds of Journal articles, probe data spanning tens of thousands of buyer conversations — and most of it lives behind a DOI, a PDF, or a paragraph three scrolls down in a paper written for other researchers, not for someone on a deadline. Answering "where's that from" by email, one reporter at a time, doesn't scale. So we pointed our own tool at the problem.
AIVO Intelligence is a citation-first conversational interface over the full AIVO Standard and AIVO Journal corpus. Ask it a question, and it answers by tracing back to the original source — DOI and all. It won't tell you what to write. It will tell you where a claim comes from, what the methodology behind it actually was, and whether a given number is completed research or a hypothesis still being tested.
That last distinction matters more than it might sound. Some of what we publish is finished empirical work — a stated method, a defined probe corpus, a result. Some of it is a working hypothesis we've proposed and haven't yet tested against data. We try to be careful about which is which in the paper itself, but we know that distinction can get lost by the time a number travels from a working paper into a headline. If a tool can trace a claim back to its actual evidentiary status in seconds, that's a small thing that avoids a real problem.
What it's for
Three use cases come up most often:
Fact-checking a specific figure before it goes in a story. If a source or press release cites an "AI visibility" statistic and you want to know what's actually behind it — sample size, method, whether it's peer-screened or a working hypothesis — ask directly and follow the citation back.
Getting oriented on a trend fast. AI search terminology is a mess right now — GEO, AEO, LLMO, AIVO, ARQ — and most of it gets used loosely. If you're parachuting into a story and need the actual distinctions between these terms and what each one does and doesn't measure, this is faster than reading four vendor blog posts that each define things differently.
Stress-testing a claim. If a company (including us) makes a claim in a pitch, ask the tool to find the tension — where does this claim hold up against our own data, and where might it be overstated.
What it isn't
It's not a replacement for talking to a person. It doesn't do off-the-record color, it doesn't have opinions about which brands are "good," and it isn't a substitute for a real diagnostic engagement if that's what a story actually needs. For anything live — a breaking story, an embargoed piece, a direct quote — email us. We aim to respond same-day, on or off record.
But for the fast, specific, "can someone just tell me where this number came from" moment that happens on every deadline, it's now sitting there, cited and ready.
AIVO Journal is the editorial publication of AIVO Standard™, publishing ongoing research into how AI systems cite, recommend, and displace brands at the point of purchase decision. For press inquiries → tim@aivostandard.org