The trust engine

It's tempting to believe AI has superpowers. It seems like it can effortlessly solve complex puzzles and like it knows everything.

But here's the deal: AI will always struggle to tell the difference between real, trustworthy information and stuff that's just not true. It's not that AI is bad – it's just not reading our minds.

Imagine you're in the kitchen, following a family spaghetti recipe that's been passed down for generations. You trust that recipe to make a delicious meal because those exact ingredients, steps, and measurements have been proven time and time again to deliver that delicious spaghetti.  It has been validated by your great-grandmother, your grandmother, your mother and today, you, without hesitation, you trust it.  

When it comes to AI suggesting a recipe, you have no way of knowing if it was imagined or based on something people actually tried and tasted. And if it's an original, how does AI even know what spaghetti actually is and what goes well with it?

Alright, quick funny story: when I was a little kid and was left alone in the kitchen I used to mess up trying to cook something delicious. I had this theory: something that tastes good + something that tastes good= Something delicious. Soon enough, after a chorizo and Nutella sandwich I realized that the formula was not quite reliable. Today AI does something similar, and surely, sometimes it works, but sometimes, it doesn't, just like that chorizo and Nutella sandwich.

Trust, the new currency

So, let’s imagine what it would look like if you could leverage the right trust engine, coupled with AI to be able to access instantly the sea of information living in your knowledge base.

It sounds simple, and we are making it simple:

• Doc Verification: This is your trust engine. A system which helps you mark your doc as verified or outdated. This is especially useful to warn your teammates of irrelevant information, and, at the same time, it lets our AI know to not consider this document a reliable source of truth.

• Trusted AI answers: Sifting only through verified docs to get the answers to your Ask request is the key to this AI-powered feature. You can sit back and trust the information you’re getting is accurate and up-to-date, just like your family’s spaghetti recipe.

The future? An AI assistant detecting docs on your behalf that need to be deprecated, and the ones that need to be verified.

We're already taking the first steps towards building a knowledge base on autopilot that teams can rely on, and we've we’ve got exciting news coming up soon so keep an eye on our changelog, and follow us on linkedin, twitter, and producthunt to stay in the know. You won’t want to miss out.