Practiced a mock interview with an AI tool before the real thing and it caught a rambling habit no human ever pointed out

Was skeptical going in, figured it'd just be a chatbot asking generic questions. Ran through a couple of behavioral prompts and system design ones, and reading back the transcript afterward I noticed I kept trailing off into three-sentence tangents before actually answering the question. No mock partner or friend had ever called that out, probably too polite. Tightened up my answers before the real interview and it went noticeably smoother.