Citation Ledger tracks whether AI engines name your brand when buyers ask, fixes the reasons they don't, and hands you a scoreboard that updates every month. Every claim sourced. Every limit disclosed.
1,125 engine observations per client, per month. Raw data ships with every report.
A buyer using AI doesn't see pages of results. They ask ChatGPT, Gemini, Perplexity, or Claude and get an answer that names three to five brands. 35% of US consumers already start product discovery with AI tools, versus 13.6% who start with traditional search.[4]
The traffic behind those answers is small but compounding: AI referrals crossed 1% of all website traffic and are growing roughly a percentage point per month[5] — 1.13 billion referral visits in a single month last year, up 357% year over year.[6] And the visitors who do click arrive pre-qualified: they've already seen you compared against your competitors inside the answer.
Which produces the uncomfortable math: if the engines don't name you, you were never even an option.
See the full data on where search is headed →
The channel most marketing budgets were built on is quietly paying out less every quarter. 58.5% of US Google searches now end without a click to any website[10] — and when an AI Overview appears above the results, users click a traditional listing only 8% of the time, versus 15% without one.[11] Ahrefs measured the #1 organic position losing 58% of its clicks when an Overview sits on top of it.[12] Ranking first now often means being the best-labeled door on a hallway nobody walks down.
Marketers are feeling it in the reports: 51% say their organic traffic declined in 2025 because of AI search.[13] Meanwhile the buyers didn't stop researching — they moved.
42% of B2B buyers used an AI assistant to research a vendor in the past 90 days.[14] The demand is intact. The doorway moved.And here's the part that decides winners: this isn't a uniform decline. Brands that are cited inside AI Overviews earn 35% more organic clicks than uncited competitors on the same queries.[15] The answer layer doesn't shrink the pie evenly — it hands the pie to whoever's named in the answer.
If you found Citation Ledger because ChatGPT, Claude, Perplexity, or an AI Overview named us when you asked about getting your brand into AI answers — yes. We know exactly how convenient that looks. It’s also the entire pitch, demonstrated: we run this program on ourselves first, with the same prompts, the same engines, and the same monthly ledger we’d run for you. Consider the referral our proof of work.
And if you got here through Google: the old doorway still works. That’s rather the point of the page below.
The finding that changes the plan: when we map where engines source their answers in a category, the client's own website is rarely the main input. The engines lean on a short list of third-party authorities — review platforms, comparison articles, communities, Wikipedia, a couple of trade publications.
That reframes the work. Getting cited isn't only about fixing your site. It's about earning your way into the sources the engines already trust. That's the part of the program that takes months, not weeks — and it's why this is a retainer, not a report.
Citations remain rare and rising: the share of ChatGPT answers carrying visible citations grew from 0.6% to 2.8% in eight months[8] — which is exactly why we track unlinked brand mentions alongside linked citations. Mentions are most of your exposure. We report both, separately.
Every tier gets the identical program — the full protocol, monthly re-runs, and execution. Pricing follows your annual revenue, not a feature ladder. Six-month initial term, month-to-month after.
The Ledger Guarantee: if your end-of-term re-measurement doesn’t show clear improvement in mention share over your approved baseline — same prompts, same engines, same protocol — we refund your program fees. Three conditions, published in full.