AI agents now tell the truth about what they've done, and stop retrying when something fails

Created documents open automatically in the chat When an AI agent creates an intake form or job description, the document now opens right beside the chat as soon as it's ready. Clicking a link like "Open intake form" takes you there without reloading the page, so your conversation stays open.

AI agents no longer claim they've done things they haven't Agents could tell you an intake form or job description had been created when nothing actually existed. Agents now only confirm actions they've completed — and the content they generate reflects everything you asked for, not just the job title.
AI agents stop retrying when a step fails When an agent couldn't complete a request, it was silently retrying the same step over and over. It now tells you once that something went wrong and stops, so you're not left waiting on a loop that was never going to succeed.