🚀Update: We gave our AI a "Brain Dump" feature!
We shipped something this week based on what you kept telling us.
Most AI tools make you jump through hoops to teach them anything. Upload a document. Format a spreadsheet. Follow the schema. By the time you're done, the flash sale is over.
We built the opposite.
Brain Dump. Open your dashboard, type whatever you need the AI to know, hit save. That's it. "Give anyone who asks a 20% discount with code FLASH20" becomes an active rule in seconds. No formatting. No uploads. Just type and go.
Teach AI. Reviewing a live chat and the AI said something slightly off? Click the response, edit it into what it should have said, save it. That correction becomes a permanent rule. The AI never makes that mistake again. Your whole support operation gets smarter every time you review a transcript.
New chat widget. We killed the FAQ tabs. Nobody wants to scroll a list when they have a question. The new widget is a clean, conversational interface. Customers ask, the AI answers. It looks premium because it actually is.
Three features, one idea: your AI should bend to your business, not the other way around.
Try Brain Dump today, especially if you run any kind of time-sensitive promotions. We want to hear how fast it takes you from idea to deployed.
Drop your use case in the comments. What rule would you inject first?
Replies
@joseph_akhatasebhudo This feels useful because support teams don’t always have time to “train” AI properly. Sometimes you just need to type one quick rule like “use this discount code today” and move on. The transcript correction loop also sounds practical — that’s where a lot of small support mistakes can actually get fixed.
@alpertayfurr Exactly the thinking. No time to sit and train an AI, just drop a rule and move on. Same with the correction loop, fixing mistakes as you go beats trying to get it perfect upfront.
@joseph_akhatasebhudo Exactly. Fixing it while the real work happens feels much more natural than trying to train everything upfront.
hey joseph, this feature answers my exact concern from yesterday. brain dump + teach AI together is the loop most support tools never close. one is "AI knows what to do," the other is "humans can correct what it did wrong without rebuilding the training pipeline."
a quick one: when a teach correction conflicts with an existing brain dump rule, who wins? if brain dump says "give 20% with FLASH20" and a teach correction says "no, it's 15% now," does the latest win automatically or does it flag a conflict?
asking because that's the moment most rules engines turn into "why did the AI say something we explicitly told it not to say" tickets.
@thenameisarian Hey Mustafa, glad it's clicking! No conflict flag the Teach Correction just wins automatically. It loads after the Brain Dump, so the AI naturally follows the more recent instruction. Only edge case is two conflicting corrections, but the admin can clean those up from the dashboard.
@joseph_akhatasebhudo clean answer. "latest wins + dashboard surfaces conflicts for cleanup" is exactly the right escape valve.
one thing i'd consider for v2: an "active rules" view that shows what the AI thinks the current ground truth is. when you have 40+ rules layered on top of each other it gets hard to predict what the AI will say without simulating the conversation. just a thought.
@thenameisarian That's a sharp catch. Past a handful of rules, conflict resolution isn't enough, you need visibility into what's actually winning right now. Adding that to the list for v2.
@joseph_akhatasebhudo appreciate that. the meta thing i keep noticing in rule based systems is that "which rule fired here" visibility beats trying to predict in advance. happy to be a beta tester for v2 if useful.
The "Brain Dump" concept is incredibly smart. I'm currently building a native Telegram Mini App for finance, and managing AI behavior without redeploying code is a real pain.
If I were using this for support, the first rule I'd inject would be: "If a user asks why their Telegram Stars payment failed, always tell them to check their local App Store limits first before escalating to human support."
Giving non-technical founders or support teams the ability to just type that in plain text instead of updating a JSON schema is a massive time-saver. Congrats on shipping this!
@alex_dev_user Yeah that's literally the whole point of Brain Dump. Knowing the answer shouldn't be the easy part while getting it live is the hard part. Love the Telegram Stars example too, that's such a real one. Good luck with the Mini App man, that space is wide open right now.