Launching today
Maskin
AI made individuals faster. Maskin makes companies smarter.
15 followers
AI made individuals faster. Maskin makes companies smarter.
15 followers
Build the loops that make your company AI-native. From customer signal to shipped bet to measured outcome — automatically. Open source.
Free Options
Launch Team / Built With
Sure
Hi Product Hunt Community,
Two months ago Magnus and I made a bet. Can we build the workspace for AI Agent & Human collaboration using 3 structured elements: Bets, insights and tasks. Today that Machine (Maskin is Scandinavian for Machine) is running ours entire company using closed loops.
Magnus and I trained PMs in AI PM before this. Every team hit the same wall. Individuals got AI but the organisations did not. ChatGPT made one person 2x faster. The company they work in operates the same as it did in 2023.
The missing layer is loops, not features. A team forms a bet — a shaped, measurable outcome. Agents do the legwork: pull the signal, draft the work, surface the result. The human reads the outcome, the next bet is sharper because of it. That's what closing the loop looks like when you describe behavior instead of selling magic.
Maskin is the workspace that shape needs. One surface humans and agents share — same objects, same comments, same history. MCP-native end to end, Apache 2.0 from day one.
Humans do the thinking. Agents do the work.
One open question: when you've worked with an agent inside a real workflow, what was the smallest thing that made it suddenly feel like a teammate instead of a tool?
PageAI
@krumhausen this looks awesome, congrats on the launch! 🥳
Persuasive Patterns Card Deck (4th gen)
Love the focus on shared objects and closed loops!
What’s the first workflow where teams usually realize Maskin is changing the company, not just making one person faster?
Sure
@toxboe Thank you! And yes, that framing nails it.
The closed loop that resonates most when we present is the customer feedback loop: agents watch Slack, Intercom, and everywhere else feedback lives, cluster the insight, and promote it to a bet the team can review. If it's a bug, it's fast-tracked straight to production
-> When the fix ships, the feedback agent notifies the customer directly. No human relaying it, no dropped thread.
We're still pre-PMF ourselves, so we run the same loop on strategy: agents read the latest AI-agent research and turn it into proposed bets that shape where the product goes next.
That's the moment it stops being "one person got faster" - the loop closes without anyone holding it together by hand and it's transparent to the entire org.