The Bouncer Method
The method separates observed answers, source associations, access diagnostics, prescription, and later measurement. A Sprint samples ChatGPT and Perplexity, records which brands and sources appear, checks access and retrieval context, and converts the strongest findings into Source Action Cards. Later execution and remeasurement are separate.
It is the same four moves on every engagement, on purpose: a repeatable process is what keeps the scope, the timeline, and the price fixed. Here is each move, and what you can hand a skeptical stakeholder at the end of it.
The four moves
Ask the real question
We ask the AI engines the exact questions your buyers ask, and record which brands the answer names instead of you, and which sources it cites.
Prioritize source actions
For priority sources, a Source Action Card connects the observed answer, competitor presence, likely action class, feasibility, evidence level, unknowns, and next decision. This is specific prescription, not execution, and it does not imply that acting will cause a citation.
Check access per engine
For each AI engine we send a request carrying its crawler's user-agent and record the response, before and after any fix, so a robots or header block shows up in writing. "Verify" here means the finding carries its record: the exact request is listed so you can re-run it yourself, though a later run can return a different response, since server behavior varies with IP, location, time, CDN, and WAF rules. It is not proof the real crawler reached you from its own IP, which only your server logs confirm.
Measure again when separately scoped
Repeated measurement belongs to public research or a separately scoped future engagement. It is not included in a single Sprint. When a panel is rerun, gaps and failures remain visible.
Why the same four moves every time. A repeatable system is what lets us quote a fixed scope, a fixed timeline, and a fixed price, and it is what makes the result auditable rather than a story. The diagnosis names what is broken; the prescribe, verify, and measure loop scopes the fix and records what changed and when, on a meter you can check yourself. It shows the sequence and the work that preceded a change; it does not claim timing alone proves the cause.
An honest boundary
The engines decide which brand gets cited; we do not. What the method works on are the inputs that influence citability: the sources the answer draws from, the content worth citing, and the access that lets a crawler read you at all. The placement engine remains under public test until the first demonstrated citation flip. We publish every dated run of our own tracked citation number, including while it is zero; the latest published row is dated 23 July 2026 and sampling is currently paused. A readiness score measures whether AI can read a site, not whether AI cites it, and we keep the two numbers separate on purpose.
Run on our own site first
These same four moves run on Deep Ocean's own domain first, signed by name: measured on a dated public ledger, published whether the number flatters us or not. See the tracked result on The AI Citation Index, or the same gap sampled on two other strong brands in the field sample.
The Visibility Sprint covers Diagnose and Prescribe through the Source Action Card, plus access findings inside Verify. Post-intervention verification and recurring Measure are not included.