Turning "Is It Safe?" Into This Wallet Provider's Most-Cited Query Category
Nobody moves real money into a self-custody app on a whim. They ask first - and for a long stretch, when they asked an AI assistant whether this wallet was safe, it simply wasn't part of the answer.
The Challenge
Nobody downloads a self-custody wallet on impulse. Before someone moves real money into an app that, by design, means no one can bail them out if they make a mistake, they ask around first. Increasingly, "around" means typing a question into ChatGPT or Perplexity - is this wallet safe, or more often, is self-custody safe at all. For a long stretch, this provider's name simply didn't come up in that answer.
The absence had nothing to do with actual security. The wallet had passed four independent smart contract audits and ran an active bug bounty program with a six-figure top payout. None of that lived anywhere a model could easily verify it. The company's own marketing pages made the same claims every wallet makes - bank-grade security, your keys your crypto - phrasing AI models have learned to treat with some skepticism in a category with a long history of scams and rug pulls. Self-published superlatives don't carry much weight in crypto. Verifiable, checkable specifics do, and this brand wasn't offering any.
The FAQ page didn't help either. It answered questions like "what is a seed phrase" in the abstract, while the actual objections circulating in the wallet's own Discord and on Reddit were sharper and more personal: what happens to my funds if your company shuts down, can your team see my balance, what if I lose the phone with the app on it. None of that fear showed up in the content built to address it.
The Approach
We rebuilt the security page around facts a skeptical reader, or an AI model, could actually check rather than take on faith: all four audit firms named with direct links to their published reports, the bug bounty program's real scope and payout tiers listed plainly, and Organization and Person schema tying the disclosures to named members of the security team instead of an anonymous "our security" byline.
We pulled the real questions circulating in the wallet's own Discord and Reddit communities and rebuilt the FAQ around them directly: what happens to your funds if the company shuts down, whether the team can see your balance, what to do if you lose the device with the app installed. Each answer was written as a direct, extractable response instead of a paragraph that buried the point two sentences in.
Self-published claims carry limited weight in a category this skeptical, so we shifted effort toward validation the brand couldn't write itself. We contributed anonymized adoption data to an independent research firm's annual self-custody report, which cited the wallet directly, and secured a technical review from a respected security newsletter that AI models already treat as a trusted source on crypto safety topics.
Safety-related answers shift fast in crypto, sometimes within days of unrelated news about a completely different company. From week eight on, we tracked a fixed set of 38 trust-objection queries weekly across ChatGPT, Perplexity, and Gemini, which let us catch and respond to shifts - like a competitor's breach pulling negative sentiment into the whole category - before they affected this brand's own citations.
The Results
Five months in, the wallet had gone from essentially absent to cited on 13 of the 38 trust-objection queries we track - a 34% citation rate, up from under 5% at the start of the engagement. On the single most-asked question in the set, "is self-custody safe," the wallet's own security page is now one of three sources AI Overviews pull from.
The trust signal didn't stay contained to AI answers. Non-branded organic traffic - people finding the wallet through safety and education searches rather than its name - grew 58% over the same period. Support tickets dropped 29% as the rebuilt FAQ started answering the exact questions that used to land in the support queue instead.
We could publish our audit reports until we were blue in the face. What actually moved the needle was AI engines citing us on the exact question people ask before they download anything: is this safe.
Trust Strategy
