In partnership with
New AI tools launch every day. A few will change how you work. The rest will be forgotten by next month.
TLDR AI is the free newsletter that tells you which is which. Every morning, Anthropic and ex-Google engineers pick the new tools, updates, and open source releases worth trying, with a quick note on what each one does and who it's for.
You find the good ones early, without testing all of them.
One email, 5 minutes a day, read by 1.1M+ people. Get tomorrow's picks.
One brand. Eight weeks. Twenty-two out of a hundred to seventy-four.
That is a single case study, not a benchmark, and we are going to keep saying so. But it is instructive about sequence. Nothing exotic happened here. Structural fixes first, answer-first rewrites second, off-site citation work third. In that order, because that is the order in which the effort pays.
The reason it moved that fast is that the structural layer was almost entirely unaddressed, which is true of most B2B sites. The cheap wins were all still sitting there.
Single brand. Not a benchmark.
Why it compounds
1 |
Content Structure → Extractability → More Citations Answer-first content makes pages extractable. Each citation is a data point AI uses to calibrate your brand's topical authority. Improve structure once. Benefit compounds as AI systems continue encountering and citing the same structured content across different buyer queries. The structure investment is a one-time fix that pays indefinitely. |
2 |
Schema → Entity Recognition → Confident Citations
Organization schema establishes your entity. FAQPage maps your brand to specific buyer questions. HowTo positions you for implementation queries. Each schema type adds a new citation pathway. Entity recognition scores accumulate over time, and a brand with 18 months of consistent entity signals is dramatically harder to displace than one with 2 months. |
3 | Third-Party Citations → Trust Signals → Tier Upgrades Every G2 review, press mention, and Reddit thread is a corroboration node. Trust signals accumulate. A brand with 2 years of G2 reviews and press coverage has a trust moat that a newcomer cannot replicate in 30 days, even if they suddenly decide to prioritize AEO. The time advantage of starting now is measured in years, not months. |
4 | Citation Frequency → Brand Recall → Direct Search Buyers who see your brand cited in AI answers remember your name, even when they don't click. That brand recall converts to direct searches 1–3 weeks later. Direct traffic is a domain authority signal. Higher domain authority means AI crawlers visit your site more frequently. More frequent crawls mean freshness updates take effect faster. The channel is non-linear: more citations → more recall → faster indexing → more citations. |
- Measurement: 20-query tracking set, weekly citation score, month-over-month trend data ready for leadership reports
- Content structure: Answer-first format on key pages, FAQ blocks deployed, BLUF headings on top posts
- Schema coverage: Organization + sameAs, FAQPage, Article with dateModified, HowTo on process content
- Brand declaration: llms.txt published, canonical description consistent across all profiles
- Third-party stack: G2 optimized with 20+ reviews, Reddit presence started, Wikidata entry created
- Freshness system: Quarterly audit running, dateModified schema updated, visible dates on key posts
The honest part
The first four weeks look like nothing is happening. Structural work does not produce citations, it produces the conditions for citations. Teams give up here, right before the curve turns.
Then the off-site work lands, and each citation makes the next one likelier, because the engine now has corroboration it did not have before. That is the flywheel. It is slow, then it is not.
 | Rand Fishkin SparkToro · on zero-click search |
Frequently Asked Questions
How long does AEO take to work?
What is the AEO flywheel?
Why do most AEO programmes stall?
Have you run this and measured it?
This newsletter is trying to work out whether AI visibility is a real channel or an expensive story. That gets settled by operators, not vendors. If you have run a programme and have numbers, good or bad, we want the case study. The ones that did not work are worth as much as the ones that did.
Book 15 minutes →A 15-minute walkthrough, no pitch. cal.com/thrivestack/citedby-ai-visibility