# A public video people and agents can actually read

Appraiser Marketing Group published [Making video content readable to people and AI](https://appraisermarketinggroup.com/videos/open-video-publishing-pilot/) to demonstrate a practical client service: turn a business's useful answers into a source customers can watch, read, verify and share.

A video link alone does not ensure an agent receives the spoken words. It can receive a title, description or error instead. Our example puts the complete transcript directly on AMG's own public watch page, alongside a playable video, timed captions and ordinary text links. The business retains a clear source for its knowledge instead of depending entirely on another platform's transcript interface.

The pilot is an original **71.708-second AMG video with disclosed AI narration**, a 178-word approved script and 18 caption cues. It demonstrates publication and access; it is not a client testimonial or a claim of professional appraisal expertise. [Watch the example and read its complete transcript](https://appraisermarketinggroup.com/videos/open-video-publishing-pilot/).

## What worked on the real public site

Tests ran on **November 15, 2026, approximately 8:07–8:10 p.m. Pacific Standard Time**, or November 16, 04:07–04:10 UTC.

| What we checked | What happened |
|---|---|
| Complete words without sign-in | Ordinary public requests returned the entire approved script. The HTML transcript, TXT, VTT and SRT agreed after whitespace normalization. |
| Actual viewing | Chrome decoded and played the video, visibly displayed English captions on desktop and in mobile-width emulation, and responded to Space for play/pause. |
| Reading without scripts | All four transcript sections remained available when JavaScript was disabled. |
| Discovery from the homepage | A fresh agent starting with only AMG's homepage found the ordinary public video link in the homepage HTML and read all 178 words through ordinary public requests. |
| Alternative text routes | Complete [plain text](https://appraisermarketinggroup.com/videos/open-video-publishing-pilot/media/2026-11-15/transcript.txt) and [Markdown](https://appraisermarketinggroup.com/videos/open-video-publishing-pilot/media/2026-11-15/transcript.md) supplied the speech independently of video playback. |

This is the immediate value: the business's actual published answer is available as a complete source, with its identity attached. A person can watch or read. An agent with a working public-request route can obtain the words without a special YouTube caption extractor or an account.

## The remaining barriers are visible too

The general-purpose agent web reader produced mixed results. One test initially failed to open the watch page and text file directly, then obtained the complete page, TXT and Markdown through their actual links. A separate fresh-context agent still encountered reader failures, while ordinary public requests to the same sources succeeded. The precise cause of that difference was not isolated. The web reader also rejected VTT's format, even though those captions worked in the browser and downloaded correctly.

Search is another step. At the recorded test time, exact-title, interior-phrase and customer-question searches did not return this new pilot. That is a dated observation, not proof that every search engine has excluded the page. Public availability, search discovery and being selected for an AI answer are different outcomes.

AMG's [robots policy](https://appraisermarketinggroup.com/robots.txt) and [AI-use policy](https://appraisermarketinggroup.com/ai.txt) permit public reading and training. This records the publisher's permission; it does not show that a provider has crawled the page or trained a model on it.

## What this means for a client

The proposed service starts with the client's real questions and expertise. We produce the video, review its captions and transcript, publish the complete answer with clear identity and sources, and maintain the page and files. A relevant service and contact path connects the explanation to the business.

Then we measure what happens: whether people and supported agent routes can retrieve the source, whether search finds it, whether AI answers cite it accurately, and whether visits become qualified inquiries. The pilot establishes a working public example and exposes specific remaining reader barriers. It does not establish lead growth or return on investment. Its bounded browser checks are not a screen-reader audit, and complete text agreement is not a claim that every spoken word received human listening review.

The deliverable is a maintained, usable publication of the client's knowledge. Business results earn their place through subsequent evidence.
