Plain-language conversations on how AI actually helps corrosion engineers and teams — from getting more out of chatbots to dedicated AI software and agents that do real work. Free, no login. From CorroZone.Full courses here: https://corrozone.com/courses#udemy
Reading the feed…
Full courses here: https://corrozone.com/courses#udemy Most people use AI like a vending machine: type a question, read the answer, walk away. This episode lays out a simple series of levels — from just asking, to refining, to uploading your own files, to tools, to agents that act on your computer, and finally to coding agents — so you can see how much more AI can do once you move past the first level. The single takeaway: find the level you're standing on, and try the next one up. Who this is for Beginners who have used ChatGPT a few times and want a plain, honest picture of how to get more o
Full courses here: https://corrozone.com/courses#udemy A chatbot hands you a fluent, confident answer in seconds — but you often can't tell what it actually did to get there. This episode builds the picture up one layer at a time: a raw large language model mainly generates text; tools let it fetch facts and run real calculations; a modern chatbot wraps the model, tools, and memory behind one seamless box; and dedicated AI software embeds that same model inside a controlled, purpose-built process that can validate inputs, run real checks, keep intermediate files, and produce a report you can t
Full courses here: https://corrozone.com/courses#udemy AI in corrosion is best pictured not as a robot taking the engineer's chair, but as a multiplier that lets the same team look at more of the evidence per engineer-hour. Inside a controlled workflow — a "harness" — it can take on routine, bounded first-pass work like organising messy notes, comparing reports, and surfacing the unusual cases, while engineers keep interpretation, responsibility, uncertainty, exceptions, and high-consequence decisions. The one rule that holds it together: the engineer stays in charge. A practical first move is
Full courses here: https://corrozone.com/courses#udemy When corrosion or materials test results vary across specimens and won't settle into one clean story, the instinct is to ask an AI for "the answer." This episode makes the case that AI's real value is the opposite: helping a team see the questions worth asking before it commits to an interpretation — widening the investigation before narrowing the conclusion. It walks four kinds of AI-assisted question — what else could explain this, what supports or weakens each explanation, which cases don't fit, and what's still missing — and is careful
Full courses here: https://corrozone.com/courses#udemy This episode was produced by an AI workflow — script, voices, and artwork — and it uses itself as the worked example. The interesting part isn't that it's AI-made; it's how. A workflow breaks a big job into steps, lets a machine do the heavy, time-consuming work of each, and — crucially — leaves behind an intermediate artifact at every step: a brief, an outline, a script, each one a document a human can open and read. Those artifacts are where the human takes over. You can inspect what the machine decided (a glass box, not a black box), de
