About these guides
This site collects step-by-step walkthroughs of features on RRX.com. Each demo follows a single end-user journey — registering an account, placing an order, applying a quote — from the first click through to confirmation, with a screenshot at every step. Everything here is 100% AI generated, end to end — the journeys were planned, driven in a live browser, screenshotted, written up, and published entirely by AI. None of the content anywhere on this site is manually created.
Have feedback or a suggestion? Visit the Help and feedback page.
These pages document a pre-production (UAT) environment. Screenshots may contain test accounts, test part numbers, and prices that do not reflect production data.
How each demo is organized
Every feature has a single page: a plain-language walkthrough of the journey, written for an end user, with a screenshot at every step. Use it to learn the feature, to explain it to someone else, or to follow along in the live site.
How they're grouped
Demos are filed into feature categories, which is what you see in the sidebar and on the home page. Each demo belongs to exactly one category, and both the order of the categories and the order of the demos inside them are set deliberately rather than alphabetically, so a journey generally appears near the ones you would read alongside it. A category with no demos in it yet is hidden until its first demo arrives. Open a category in the sidebar to see everything it contains, or use the feature guides index for the full list.
How they're produced
Every demo is produced by GitHub Copilot running the create-guide skill against a
single user journey. For each journey, the skill:
- Plans the journey's steps.
- Drives a real browser through those steps on the UAT site, capturing a screenshot at every step.
- Writes the plain-language walkthrough from what actually happened during the run.
No step is performed by hand — the plan, the browser run, the screenshots, and the write-up all come out of that single automated pass. Immediately before each screenshot is taken, the page on screen is rewritten so that anything sensitive is replaced by a fictional stand-in, which is why the screenshots are safe to publish exactly as captured.
Each journey produces a folder containing the end-user walkthrough, the step-by-step instructions the run followed, and the screenshots. Only the end-user walkthrough and its screenshots are published here. Test results are a separate exercise, written elsewhere by a separate skill, and never appear on this site.
How they're masked
The journeys are run from a real UAT account, so what appears on screen during a run includes customer names and addresses, account numbers, personal names and email addresses, and exact prices. None of that is published — and none of it is ever in the screenshots to begin with. Just before each capture, the live page is rewritten so every sensitive value is swapped for a fictional stand-in, rendered in the real value's own typography and layout. The image is clean at the moment it is taken, so a guide looks like a working page rather than a page covered in black bars.
Stand-ins are consistent: the same real value always becomes the same fictional one throughout a journey, and prices keep their magnitude, so tables and totals still read sensibly from one screenshot to the next.
Look at any price in a screenshot — $10.55, $98.55, $1,204.55. The fifty-five cents
is the tell: it means the number is fictional. Because the stand-ins are drawn from a
fixed set, the figures won't add up the way they would on the live site. The one exception
is the Difference: saving on the cross-reference guide, which is a genuine comparison
rather than a price and is published as-is.
Details that carry no personal information are deliberately left intact, because they help you follow along: quote and order numbers, part and catalog numbers, warehouse codes, dates, and city/state values. Over-masking is treated as a defect too.
What counts as sensitive is defined once, in a single shared rules file, so the masking that runs in the browser and the safety check that runs at publish time cannot drift apart. That check scans every generated page, the home page data, and every published screenshot at the end of each import, and reports anything it finds before the content goes live. Guides captured before this approach existed were cleaned up after the fact instead, with opaque blocks painted over the finished image; that older tooling is kept only as a fallback, and those journeys have since been recaptured the current way.
Fixes always go in the source demo folder, followed by a re-import — never in the generated page.
Where to start
Browse the feature guides in the sidebar, or start from the home page for a visual index.