Most AI-visibility advice is a guess. This sequence isn't.
A one-time "AI visibility audit" isn't a foundation — it's a snapshot of noise. This is the order I run, the five stages that actually move AI search, and the parts of this industry I refuse to sell you.
What everyone sells. What actually happens.
A single AI response carries almost no signal — independent research found under a 1-in-100 chance two runs return the same brand list. I track the trend across dozens of prompts and models, never a one-off score.
97% of llms.txt files are never requested by a single crawler, across 137,000 domains studied. A controlled study found schema markup's effect on AI citations statistically indistinguishable from zero. I don't bill either as AI-citation work.
No independent study shows a durable lift on that timeline, net of brand size. I'll tell you what's proven and what's a bet. Nothing else.
The sequence
Five stages, run in order. Each one has to clear before the next one starts — skip one and everything built after it sits on nothing.
Access & indexation
Confirm AI crawlers can reach your site and that you’re indexed where each engine retrieves from — Bing feeds ChatGPT and Copilot, Google feeds AI Overviews. A gate, not a growth lever. Block it silently and everything after this disappears.
Mandatory starting pointEntity consistency
Make your brand one clean, disambiguated entity — Wikidata, consistent naming, review-platform profiles (G2, Capterra). This is what lets a model attach any signal to the right company.
Content extractability
Restructure your pages so a model can quote them directly — direct answers, self-contained sections, original data. The only on-site lever with real research behind it.
Off-site & earned presence
Reviews, third-party roundups, genuine press coverage, Reddit and YouTube presence. This is where growth happens — slower and more expensive than anything above it, and the lever with the highest ceiling.
Ongoing monitoring
Only once the above is in place: track how often you’re mentioned and whether you stay in the consideration set, and catch factual errors — stale pricing, wrong category — before they calcify.
Why this order
"I start with what's binary — can AI crawlers reach your site, are you indexed, does your brand resolve as one clean entity. Only once that's true do I track consensus — and I track it as a trend, never a score."
Brand size and off-site signal predict whether you get recommended — far more than anything on your own domain. Owning your site's AI presence matters, but it's a gate you clear once. The compounding work happens off-site: reviews, press, the places models already trust.
The receipts
Ways to work together
Dedicated B2B SaaS AEO/GEO programs with comparable scope — content production, earned-media outreach, ongoing monitoring — run $5,000–$11,000/mo industry-wide in 2026. At $4,500/mo, you're paying less than that floor, and working directly with the person who built this system in-house at Atlan — not an account manager reselling it to a junior team.
Most engagements start here.
Most engagements start at Foundation Sprint and grow from there. If it's the right entry point for you, tell me your domain and I'll run a free 10-minute access/indexation check before we talk price — you'll know within a day whether Stage 1 is even broken on your site.
Questions I get
AEO / GEO / SEO — built on evidence, not hype.
Is your site even reachable
by AI crawlers?
A free 10-minute access and indexation check — you'll know within a day whether the gate is even open.