Why a baseline matters
Seenu Tech works on AI visibility, so the same standard should apply to its own website: describe the starting point before claiming improvement. A baseline is a dated, inspectable record, not a launch announcement or a reassuring label in a dashboard.
This revised Week 0 note defines the measurement approach. It does not publish a new engine run, a citation percentage or a commercial result. Earlier wording described broad findings without linking the corresponding raw responses. Those statements should not be treated as a reproducible measurement. A useful public record must make clear what was observed and what still needs evidence.
Keep two different questions separate
The technical question is whether a page can be reached and understood: its response, crawl controls, canonical address, visible content, internal links and relevant structured data. Each check needs a URL, a date and an observed result. A reachable sitemap alone does not establish that every page is indexed.
The answer question is what a particular AI tool says in response to a particular question. That requires the exact prompt, language, tool, date, answer and cited sources where available. Technical readiness does not prove answer inclusion, and a mention in one response does not validate the whole website. Record these as separate observations.
What an inspectable record should contain
For a page check, save the page path and the specific issue or successful condition. For an answer check, retain the response alongside its context. State whether the business was mentioned, whether a source was linked and whether the description was accurate. These are different outcomes.
If you later report a rate, define the denominator and the inclusion rule. A collection of selected favorable screenshots is not a representative baseline. Missing captures, changing prompts or unavailable tools should be noted rather than silently filled in. This article supplies a protocol, not a substitute for those records.
Five prompt groups to organize observations
Brand questions ask whether the business is described accurately, for example, 'What is Seenu Tech?' Category questions ask about the service category. Buyer-intent questions ask for help with a concrete need, such as a New Jersey business investigating AI search. Industry questions add the relevant business context. Expansion questions examine a market or topic the company is considering.
Use equivalent English and Korean questions where both audiences matter. Keep the wording and context stable enough to compare, and record deliberate changes. These groups organize a review; they do not prove that any prompt produces demand or that a weekly run has already occurred.
What this baseline does not claim
There is no guaranteed position in an AI answer. Responses can vary across tools, sessions, prompts and time. One favorable answer is an observation, not a durable ranking, a lead or revenue.
This note reports no citation frequency, ranking improvement, lead volume or return on investment. Those claims would require their own evidence and a defined method. A technical pass, an editorial improvement and a commercial outcome should never be combined into one success claim merely because they occurred near one another.
The next useful measurement
Start with one important buying path and a manageable question set. Capture the initial state, record a specific page change, and repeat the observations with their limitations intact. Include unchanged or unfavorable results when describing the work.
The Living Lab and Business Blog can provide places to publish records when the supporting evidence is ready. An AI Visibility Audit can apply the same distinction between technical checks, answer observations and next actions to a business. The objective is a factual starting point for decisions, not an impressive number invented to launch a series.