Every assistant we tried had the same flaw. It was clever for one conversation, then it forgot us. We explained the same job, the same people and the same constraints again and again, to something that was supposed to be helping.
The useful people in our lives don’t work like that. A good friend remembers what you said in March. A good colleague spots the clash before you do, and checks before sending something in your name. They hold the thread.
So that’s what we set out to build. Iris starts from memory rather than bolting it on: one continuing relationship that becomes more useful the more you use it. It thinks a problem through before it acts, does the follow-up with the tools you already use, and keeps working after you’ve gone.
None of that matters unless you can trust it, so we built the controls first. You can read everything Iris remembers. Nothing leaves your hands without your say-so. Your conversations are never sold or used to train AI.
If you use Iris, tell us what it gets wrong at iris@ferrolo.com. We read every message.