AI Search Visibility: Source Strategy

Priorities, in order

1. Send Ami the competitors to include in the comparison guide. We write it next week.
2. Publish the two completed articles as blog posts: How to Choose a Crypto Mass Payout Provider and Payout Traceability, Reconciliation and Exception Handling.
3. Pitch The Next Web and PaymentsJournal.
4. Stand up the documentation subdomain. Full specification in the Documentation Build Guide.
5. Approach Finix and ConnectPay.
6. Send the five long shots as one batch.
7. Seed customer reviews on Gartner Peer Insights.

Ongoing: update one existing article every week. The change has to be meaningful, so new sections, revised figures or added detail, rather than a touched "last modified" date. We supply the copy.

Not doing: Reddit, G2/Capterra/Trustpilot, LinkedIn, mass listicle outreach.

Rough effort split: 70% own content, 20% earned media, 10% outreach and niche placements.

The reasoning behind each of these follows below.

Summary

• Nobody owns this space. Across 50 buying questions and five assistants, your named competitors were cited a handful of times between them. The answers are currently being written by adjacent fintechs who simply published better explainers.
• ChatGPT, Claude and Gemini read three different internets. No single campaign moves all three. ChatGPT in particular ignores marketing content entirely and reads regulators and technical documentation.
• The main lever is your own content, not outreach. Three properly built pages on your own domain will do more than any placement campaign, and it is the only thing that affects ChatGPT.
• A small number of external placements are worth pursuing. Two earned-media pitches and two comparison listings, detailed below.
• We are not recommending Reddit, review sites, LinkedIn or a listicle mailshot. All four were tested against the data and none of them work here. Reasoning included, because these are the obvious plays and it matters that they were ruled out on evidence.

The main lever: your own content

The content programme:

Publish the buyer's guide first. Publish the traceability reference about three to four days later. Spacing them establishes a publishing cadence rather than a one-off burst, and it lets you catch any technical problem on the first page before it is repeated on the second.

Where we should appear

Tier 1: Earned media (2 pitches)

Both convert at 100%: every time an assistant retrieved these pages, it used them.

The strongest single page in the entire set is that TNW article, and its subject is precisely your positioning. A contributed piece or expert commentary here is the highest-leverage external action available.

Also worth a low-cost approach: Bitcoin Foundation, non-commercial, high trust, and cited every time retrieved.

Tier 2: Comparison listings (2 confirmed)

The pitch in both cases: per-recipient blockchain transaction hashes as native proof of payment, and support for verticals mainstream providers decline. Neither capability appears anywhere in the current listings.

Tier 3: Long shots (5, one email each)

Finix demonstrated that payout vendors do run genuinely inclusive comparisons. These are the same format from companies whose products overlap more closely with yours. Lower odds, but cheap to try, and all are fiat-only so the crypto argument still holds.

Dots · Routable · Tipalti · HighRadius · Trolley

Expect 1 to 2 placements from all seven approaches combined.

What we recommend against

Each of these is a plausible-sounding plan that the data does not support.

No Reddit campaign

Reddit surfaced more than any other community source: 49 pages. It was almost never used. Only 2 of those 49 were cited in an answer.

For context: across the drill-down set, roughly 44% of retrieved pages get cited. Reddit converts at 4%. The models look there and then decline to rely on it for commercial recommendations.

There is a second reason. Seeded Reddit recommendations are, structurally, undisclosed promotion of a regulated crypto service. That sits badly against your Reverse Solicitation position and against a brand built on being payment infrastructure specialists rather than crypto marketers. The upside is small and the downside is not.

No G2, Capterra or Trustpilot

Combined appearances across the drill-down set: zero. Software review directories play no part in how AI assistants answer these questions. Gartner Peer Insights appeared twice and is the only review platform with any presence at all. Worth seeding reviews on, not worth a programme.

No LinkedIn campaign

Four pages retrieved. No meaningful citation activity. LinkedIn remains useful for demand generation; it is not a lever for AI visibility.

No mass listicle outreach

Thirteen articles surfaced as possible placement targets. The instinct is to email all of them. Eleven structurally cannot list you:

One pattern is worth carrying into any future outreach: "10 Best X" articles name rivals candidly; "How to Choose X" articles never do. IFX's piece is literally titled "How a bulk payment solution works (using IFX as an example)."

That leaves Finix and ConnectPay, both covered above.

The finding that shapes everything

We traced every source the five assistants consulted across the 50 buying questions, drilling into the subset carrying genuine purchase intent. Those produced 1,530 web pages. The pattern in them is the basis for every recommendation above.

ChatGPT, Claude and Gemini are reading three different internets. A single campaign cannot move all three.

ChatGPT: reads regulators, not marketing

The sources ChatGPT pulled were primary reference documents almost without exception. No blog posts, no buying guides, no trade press.

Its source set barely intersects with the other four assistants at any point.

What this means: you cannot buy, pitch or earn your way into ChatGPT's sources. There is no listicle to join. The only lever is publishing documentation-grade material on your own domain that sits credibly alongside these references, and cites them. Stripe appears here because docs.stripe.com reads like a specification, not a brochure.

Claude: searches rarely, remembers broadly

Claude ran roughly one search per question, the lowest of any assistant. It answers mostly from what it already knows rather than from live results. When it does search, it uses the mainstream fintech corpus.

What this means: Claude rewards being described consistently across the durable web over time. It is the slowest channel to move and cannot be campaigned.

Gemini: most responsive to placement

High search volume, and it draws heavily on exactly the kind of comparison articles you can realistically appear in.

What this means: Gemini is where effort converts to visibility fastest. We will use it as the early indicator that the programme is working.