<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Rpa on OpenAdapt Blog</title><link>https://blog.openadapt.ai/tags/rpa/</link><description>Recent content in Rpa on OpenAdapt Blog</description><generator>Hugo -- 0.164.0</generator><language>en-us</language><lastBuildDate>Fri, 17 Jul 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://blog.openadapt.ai/tags/rpa/index.xml" rel="self" type="application/rss+xml"/><item><title>The silent wrong write: your automation should halt instead of guessing</title><link>https://blog.openadapt.ai/posts/silent-wrong-action/</link><pubDate>Fri, 17 Jul 2026 00:00:00 +0000</pubDate><guid>https://blog.openadapt.ai/posts/silent-wrong-action/</guid><description>Screen-only verification silently passed 5 of 7 transactional fault classes — a green banner over a wrong database. We found the failure class in our own engine first, fixed five silent wrong-write modes, and built effect verification against the system of record: 55.6% silent wrong-action rate driven to 0%.</description></item></channel></rss>