📖 Self-learning AI agent to create and update help docs
Brainfish's Self-Generating Knowledge Base automatically generates help docs and delivers it exactly when needed. Like having a documentation team that works 24/7 (and reads minds). Embed a few lines of code in your product and watch the magic happen.
👋 Hi Product Hunters! I'm DK from Brainfish, and today we're excited to launch our Self-Generating Knowledge Base feature as part of our ambient AI agent.
Here's the problem we noticed: Companies spend hours creating and maintaining help content, yet users still struggle to find answers. Traditional AI solutions promise automation but require extensive setup and don't always easily understand your product's context. Meanwhile, your best support people naturally learn how users interact with your product and provide perfect guidance at the right moment. We asked: what if AI could do that?
That's exactly what we built. Our ambient AI agent works like your most attentive support person:
- It actively observes how users interact with your product
- Automatically creates help content based on real usage patterns
- Delivers contextual assistance exactly when users need it
- Continuously learns and improves without manual training
What makes this different?
- Zero setup time - Just add our snippet and the AI starts learning immediately
- True contextual understanding - It doesn't just search docs, it understands how users actually use your product
- Automatic content generation - Creates and updates help content based on real user behavior
- Natural delivery - Provides help within the user's workflow, not through intrusive chatbots or separate windows
Where does this fit in your tech stack?
Supplements & enhances:
- Traditional knowledge bases (Zendesk, Help Scout, Confluence) - by automatically creating and updating content
- Product analytics tools - by adding context to user behavior patterns
- Existing support workflows - by providing instant, contextual assistance without disrupting escalation paths
Can replace or reduce reliance on:
- Manual documentation tools
- Basic chatbots that create friction
- Standalone product tours and onboarding tools
- Multiple disconnected help content repositories
Think of it as a layer that sits on top of your existing systems, learning from everything and making it all work better together. It doesn't force you to rip and replace your current tools – instead, it makes them smarter and more connected.
Real results from our early customers:
- Smokeball achieved an 83% self-service rate without creating any friction
- One customer improved their support NPS from 60 to 77
- Another reduced training hours by 30%
Perfect for:
- Product/support teams tired of maintaining outdated documentation
- CX leaders who want to improve self-service capabilities
- Companies looking to scale support without losing quality
- Teams frustrated with complex AI tools that require a lot of setup
We built this because we believe making products easier to use shouldn't mean putting up walls between users and help. Traditional knowledge bases are static and outdated. Chatbots have become frustrating and disconnected.
Try it out and let us know what you think! We're especially interested in hearing how this compares to your current documentation and support processes.
P.S. - If you're wondering about the "video game" analogy we sometimes use: just like how modern games learn and adapt to your playstyle in real-time, our AI agent learns and adapts to how people actually use your product. No manual required! 🎮
