{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/428499f73e57490c818743d7e2de3d77\" frameborder=\"0\" width=\"3336\" height=\"2502\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":2502,"width":3336,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":2502,"thumbnail_width":3336,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/428499f73e57490c818743d7e2de3d77-3d843154d6c2a29b.gif","duration":182.046,"title":"Kane verifies AI posts before publishing","description":"This Loom demonstrates an AI agent testing and correcting social posts before they go live to avoid false promises or tracking issues. Dan describes that his AI can generate posts quickly but previously did not verify whether the content was true. He uses a tool called Kane that opens the real browser, checks every promise and verifies tracking tags, then fixes errors such as incorrect tracking tags and wrong numbers before allowing posting. He reruns the process to confirm it passes, and notes Kane has saved 39 learned fixes in the repository, adjusting older posts when the app changes."}