Guide
Why the group chat is the right place for AI
Plenty of teams try AI once, in a tab by itself, and quietly stop. That's not usually a sign the AI was wrong. It's a sign of where it lived. Here's why the group chat, not a separate tool, is where AI actually sticks around.
Why do separate-tab AI chatbots get abandoned?
Opening a new tab, remembering a login, and re-explaining your company from scratch is a real cost, even if it takes thirty seconds. Thirty seconds of friction is enough to lose most of the questions someone would have asked, because asking a colleague in the same window is free by comparison.
A tool that only one person opens also only benefits that one person. The other five people on the team never see the answer, never learn from it, and have no reason to open the tab themselves. Usage stays flat, then drops to zero.
A shared channel teaches the whole room, not one person
Glitch runs in the same chat as the rest of the team, channels and DMs, so a question asked with an @glitch mention gets an answer that streams into the conversation everyone in that channel can see. Nobody has to relay the answer secondhand.
That changes what a single question is worth. One person asking about a shipping delay teaches the whole channel what happened, not just the asker. The answer is read the way any other message in the channel is read, by everyone who's in it.
Corrections become shared memory, not a private fix
In a separate tool, a correction only fixes what one person sees. Everyone else keeps getting the old, wrong answer until they happen to catch it themselves.
In Glitch, a correction made in chat is acknowledged and saved to shared memory for the whole team. Tell it once that the vendor changed, and every future answer, to anyone who asks, reflects the correction.
Follow-ups don't need re-tagging
A follow-up question in a personal chatbot usually works fine, because the whole conversation is private and contained. In a group channel, the harder problem is knowing whether a follow-up is still talking to the AI or has moved on to a person.
Glitch follows up without re-tagging, so a second question in the same thread doesn't need a fresh @glitch mention to get answered. The conversation stays natural instead of turning into a string of tags.
Notifications make AI output part of the normal reading
An answer that lives in a tab nobody has open might as well not exist. An answer in the team's chat shows up the same way any other message does, with browser notifications and unread badges marking it as something to read.
That's the real shift. AI output stops being a separate thing you have to remember to check and becomes part of the same reading habit the team already has for everything else in chat.
Common questions
Does everyone in the channel see the AI's answers?
Yes. When Glitch answers an @glitch mention in a channel, the reply streams into that channel where everyone in it can read it, not just the person who asked.
What happens when someone corrects a wrong answer?
The correction is acknowledged in chat and saved to shared memory for the whole team, so future answers reflect it for anyone who asks, not just the person who made the correction.
Do I have to keep mentioning @glitch to continue a conversation?
No. Glitch follows up on a thread without needing to be re-tagged, so a second or third question in the same conversation still gets answered.
See it working, not described
The sandbox is a seeded company you can wander. No signup needed.