“A founder buys from a named human with receipts, not a brand account with no face.” — from the Suede launch playbook
The line above is from my own launch playbook, where it decided something small: which account posts. Everything for Suede ships from my personal account, under my name, with evidence attached. The brand account amplifies; it never originates. The reasoning was simple: a founder buys from a named human with receipts, not a brand account with no face.
I did not expect that sentence to double as technical SEO advice. Then I watched what the engines cite, and it turns out machines evaluate trust the way skeptical buyers do. This chapter is about making your pages carry receipts, because in a world of infinite generated text, receipts are what remains scarce.
Why machines rediscovered trust
Put yourself back in the engine’s position from Chapter 4, with one new pressure added: the web is now flooded with plausible, machine-written content, and citing something wrong or fake embarrasses the engine in front of its user. Every major engine is therefore tuned, and continuously re-tuned, to prefer sources that look like they know what they are talking about.
What does that look like, mechanically? It looks like E-E-A-T, the framework from Google’s own quality guidelines: Experience, Expertise, Authoritativeness, Trustworthiness. Founders hear that acronym as branding mush. Read it instead as a checklist of forgeable versus unforgeable signals. Adjectives are free and infinitely forgeable; any text generator can produce “industry-leading.” Receipts are expensive to fake: a named author who verifiably exists, a specific number with a date and a methodology, a screenshot of a real result, a client who says so in public, a document trail that third parties corroborate. Engines lean toward the expensive signals for the same reason buyers do. That is the entire theory of this chapter. Now the practice.
Receipt one: a named human
Anonymous content is the weakest content on the modern web, because anonymity is what generated filler looks like. Put names on your work. Your key pages and articles should carry a real author with a real bio: who they are, why they are qualified on this subject, where else they exist, their site, their profiles, their history. Mark it up with Person schema from Chapter 8 so the entity plumbing connects the name to the work.
This is Experience and Expertise made machine-legible, and it is also why the faceless-brand pattern fails twice: buyers do not trust it, and now neither do the machines. If you are a founder, you are sitting on an unfair advantage here. Your story, your reasons, your firsthand account of the problem you solve, none of it can be generated by a competitor’s content vendor.
Receipt two: specifics with dates and sources
Go through your key pages and find every claim shaped like “faster,” “trusted by thousands,” “best in class.” Each is a receipt-shaped hole. The repair is the evidence block from Chapter 8: the specific number, with its date, source, or methodology attached, written as an extractable passage.
“Significantly faster imports” becomes “In our March benchmark, importing a 10,000-row catalog took 41 seconds, down from 9 minutes in the previous version; methodology at the link.” One of these is quotable by a machine trying not to embarrass itself. The other is fog. And when you cite numbers that are not yours, name where they came from and when. An engine, like a careful reader, treats a sourced claim and a floating claim as different species.
The discipline has a hard edge that most marketing resists: if you do not have the receipt, do not make the claim. An unsupported superlative is not neutral filler; it is a signal about everything else on the page. Write around the gap or go earn the number.
Receipt three: other people’s words
Chapters 4 and 5 established that engines lean heavily on third-party surfaces for buying questions: review platforms, comparison articles, community threads. That is trust logic too. Your page says what you claim; the outside record says what the world corroborates. Which makes part of AI visibility work happen off your site entirely.
The playbook is patient and honest. Be present and accurate on the review platforms that matter in your category, and actually ask your happy customers to write there; most silence is unasked, not unhappy. Be genuinely useful in the communities where your buyers ask questions, which produces threads that mention you for real reasons. Publish reference material accurate enough that comparison writers use it, since they are drafting tomorrow’s citations. None of this is manufacturable in a week, and that is precisely why it works. It is the evidence flywheel from Chapter 5, deliberately spun in your direction.
The disclosure edge
Here is the counterintuitive move that ties this chapter together: disclose what promoters hide.
My own testimonial protocol says every quote from a founding-rate client ships with a disclosure line attached, verbatim, stating that the client got a reduced founding rate and that the rate was not conditioned on what they said. My delivery emails carry a point-in-time disclaimer and a refund policy on every offer. Say plainly what your product does not do; state the conditions under which your numbers were measured; put the caveat in the copy.
The reflex says this weakens the pitch. Watch what it actually does. A page that discloses reads as a page with nothing to hide, to a buyer and, increasingly, to systems tuned to detect the texture of trustworthiness. One honest limitation makes every adjacent claim more credible. In a sea of breathless generated copy, the disclosed caveat is the strongest scarcity signal you can print, and unlike every other receipt in this chapter, it costs nothing but nerve.
Check this yourself right now
Two audits, ten minutes each.
Onsite: open your most important page and count claims versus receipts. Every “leading,” “trusted,” and “fastest” goes in one column; every named author, dated number, sourced stat, and disclosed caveat in the other. Most pages run heavily claims-first. Convert two: one adjective into an evidence block, one anonymous section into a named one.
Offsite: ask Perplexity, “What do people say about [your company]? Include sources.” Read the citations. That is your third-party evidence trail as a machine sees it, and if it is thin, you have just met the patient work this chapter assigned you.