In our own open-source tracker, TypelessForm — our voice-to-form widget — was present in all 12 of 12 checks across ChatGPT, Perplexity, Gemini and Claude on 11 July 2026 (named in 11, cited in one): a 100% presence score, Unified Visibility Index 92 of 100. We built the tracker and graded ourselves — here is the full, inspectable paper trail, caveats included.

This is a case study about a single, verifiable data point, published in full so anyone can check our work. It is not a promise about traffic, sign-ups or revenue — we return to that plainly at the end. What it is: a documented run of an open-source measurement tool showing one of our own products reaching complete brand presence across the four AI answer engines we track, and the full trail of runs that led there.

Who ran this measurement — and isn't grading yourself a conflict of interest?

Yes, it is, and we want to be the first to say so. Webappski built aeo-platform, the open-source tracker used here, and TypelessForm is our own product — we are grading our own homework. The reason we can still publish it honestly is that the tool is open source: the queries, the scoring formula and the raw run are all inspectable, so you do not have to take our word for a single number.

Anyone can read the exact weighting our Unified Visibility Index uses (it lives in lib/report/visibility-index.js), install the same command-line tool from npm, point it at their own brand, and get the same kind of report. A claim you can re-run yourself is a very different thing from a marketing number you are asked to trust. That is the whole point of doing this in the open.

What does a 100% score actually mean here?

It means that in this one run, on 11 July 2026, TypelessForm appeared in every one of twelve checks — four AI engines times three category queries. Eleven of those were direct mentions inside the engine's written answer; one (Claude, on the third query) was a citation-only appearance, where our domain was among the sources but not named in the prose. The presence score counts both as present.

So the honest reading is: twelve cells, twelve hits, with one of them a weaker source-only signal rather than a full sentence. Twelve cells across three queries is a snapshot of one moment, not a verdict on the entire market. It tells us the engines can find and describe the product for the questions we asked — which is exactly what we set out to measure, and no more than that.

How did the score move from 33% to 100%?

Across seven runs between 23 April and 11 July 2026, the presence score rose from 33% to 100% — a 67-point gain. But those two endpoints are not a like-for-like comparison, and we will not pretend they are: the engine panel was not identical across all seven runs. Perplexity, in particular, was absent from the panel in some earlier runs, so the raw 33% start and the 100% finish were measured against different sets of checks.

The clean, same-panel comparison is the most recent pair the tracker can compare directly: 83% on 11 June to 100% on 11 July, counting only the cells present in both runs. In that honest window, Perplexity moved from missing two of its three queries to answering both, while Claude's third query softened from a full mention to a source-only citation. The June path itself was not a straight line — the score first touched 100% on 10 June, slipped to 83% the next day, and returned to 100% by 11 July.

One more precise word on cadence: these were seven runs spread across the roughly eleven weeks from 23 April to 11 July, not a weekly schedule and not a guarantee of weekly measurement. We run the tracker when it is useful to, and the trend is the shape of those seven runs — nothing more regular is being claimed.

What do the AI engines actually say about TypelessForm?

Across the run the engines described TypelessForm, in their own words, as a one-shot voice-to-form widget you add with a single script tag, supporting more than 25 languages including cross-language input. These are the sentences a real person reads when they ask an assistant for voice-form tools — the current "AI snippet" for the product. Three, quoted verbatim from the run:

Gemini: “TypelessForm (Best ‘One-Shot’ Form Widget) How it works: It is a highly specialized widget added via a single script tag.”

ChatGPT: “TypelessForm | One-shot voice-to-form: speak one sentence, it fills all fields. Supports 25+ languages including cross-language input (e.g., speak Spanish, form fills in English).”

Claude: “TypelessForm is positioned as the one-shot voice form filling widget for a site's existing forms, supporting 25+ languages with cross-language capability.”

Where do the AI engines get their answers?

In this run the single most-cited publisher in the whole category was typelessform.com itself — 24 citations, 16.7% of every source the engines pulled from — ahead of each competitor domain. Our agency site, webappski.com, appeared five times. That concentration matters because it shows the engines are leaning on the product's own well-structured pages, not on paid placement, to describe it.

It also points at the actual lever behind the score. Presence in AI answers did not come from advertising; it came from pages built so an engine can lift a clean, self-contained sentence out of them. That is the method we apply for clients too, and it is the reason a small product can out-cite larger competitors in a niche.

How was Perplexity measured, specifically?

One honesty note on method. Perplexity's three answers in this run were collected by hand — we pasted each query into Perplexity and recorded the response — rather than through an automated API like the other engines. That is a deliberate, disclosed limitation: Perplexity's cells reflect a manual snapshot on the day, not a scripted, identically-repeatable API call. We flag it so the run is read for exactly what it is.

Can you reproduce this yourself?

Yes — reproducibility is the reason we published it. aeo-platform is an open-source command-line tool on npm. You install it, point it at your own brand and your own category queries, and it produces the same kind of report: an engine-by-query matrix, a visibility trend, a domain share-of-voice table, and a Unified Visibility Index whose weighting — presence 35%, sentiment 25%, rank 20%, citation 20% — is written in plain code you can read, and change, before you trust it.

What this does — and does not — prove

This proves one thing precisely: on the day and the queries we tested, an open, inspectable tool recorded our product at full presence across four AI engines, with one of those twelve cells a weaker source-only signal. It does not prove sales, sign-ups, website traffic, or that we will hold the position tomorrow. AI answers shift as engines and content change; a visibility score is a signal to act on, never a finish line — for us or for anyone we measure.

Frequently asked questions

What is aeo-platform?

aeo-platform is our open-source, command-line AI-visibility tracker, published on npm. It runs a set of category queries against ChatGPT, Perplexity, Gemini and Claude, records whether a brand is mentioned, and scores the result into a report — a matrix, a trend, a share-of-voice table and a Unified Visibility Index. Because it is open source, the queries and the scoring formula can be inspected and re-run by anyone.

Does a 100% score mean TypelessForm is the top voice-form tool?

No. It means TypelessForm was present in all twelve checks we ran on 11 July 2026 — four engines, three queries — not that it wins every possible question or every market. Presence is a signal that the engines can find and describe the product for the queries we tested; it is not a verdict on the whole category, and it says nothing about sales or traffic.

Why should I trust a company that graded its own product?

You should not trust it on faith — you should check it. We disclose plainly that we built the tracker and that TypelessForm is ours. What makes the claim different from marketing is that the tool is open source: the queries, the scoring weights and the raw run are all inspectable, and you can install the same tool and re-run the measurement against any brand, including your own.

Does this measure Google rankings or website traffic?

No. This measures AI-answer presence only — whether the engines mention a brand when asked a category question. It is not a Google ranking, not a traffic count and not a conversion metric. It sits alongside your SEO and analytics, answering a narrower question: when someone asks an AI assistant, does your brand show up in the answer?

Can I get this measured for my own site?

Yes. We run the same open tracker on client sites, and the first AI-visibility check is free — you get the matrix and the honest read, not a sales pitch. If you want to see where your brand stands with ChatGPT, Perplexity, Gemini and Claude, request a free audit at webappski.com.

Want the same paper trail for your own brand?

We run this exact measurement — the open tracker, the engine-by-query matrix, the honest caveats — on real client sites, and the first AI-visibility audit is free. If you want to know whether ChatGPT, Perplexity, Gemini and Claude mention you when it matters, request a free audit at webappski.com.