The End of Endless Search
I love structure. I adore process. I delight in complex architecture. I am, at heart, a rules-based person. All of that is why, some years ago, I wrote a piece on this very website arguing that you had to think about all of these things when adding information to your knowledge base, and that the information itself had to be bite-sized in order to properly maintain and consume. It turns out I was wrong. Sorry about all that.
Which isn’t to say that I was wrong at the time. I was right! (I also love being right.) But that’s because the task of making knowledge easy to consume was one that fell to us. We had to make things simple to search for; we had to consider what was most important; we were required to think of how it all fit together. We had to be the curators of the index.
But that’s no longer true. With the advance and high adoption of generative AI, we no longer have to consider all of that. Freed from the burden of considering deep structure, we can focus on the more important stuff of sharing and learning.
How generative AI changed the knowledge game
In that 5-year-old article, I said, “If your corporate knowledge portal is like most, when you have a question, you have to know a) exactly what you’re looking for or b) you have to search through hundreds (or thousands!) of words in multi-page FAQs or PDFs to find an answer that is buried somewhere in them.”
In a separate article (“The Best Search is No Search Powered by AI”), a colleague argued something similar:
“No search is the concept of having needed information pushed to you without you having to know exactly what you're looking for. No search does not mean that search is going to disappear. Search isn’t going anywhere. We will always need search to surface things, unless someday we develop technology that can read our minds (a scary thought). When we say ‘the best search is no search,’ that’s not to advocate for eliminating search altogether. It's only to say that sometimes there is an alternative to search, and that involves discovering information without having to physically do the searching.”
Even then, we knew that traditional setups and searches were inefficient at best, and useless at worst. And while Guru was already a great solution to fixing those problems, the actual management part of knowledge management was still a big part of making the end result successful.
Generative AI has changed all of that. While it’s still critical to ensure that what you’re putting into your KB is correct (GIGO—garbage in, garbage out—is still a big factor), our genAI features like Answers prove the best search is no search. Answers checks everything, and instantly gives you a unified answer to the question you have, without you having to check search results.
The best structure is no structure
OK, that’s not entirely true. You don’t simply want to add 10,000 words to one article and call it a day. Times and information change, and you still want to make sure you can easily update what you have in there. But considering how to interlink a series of Cards or how exactly to phrase something doesn’t much matter anymore; it’s AI’s burden now.
This also extends to the actual content of your information as well. Contributing information that’s going to be widely used can be pretty intimidating regardless of how well-informed you are. Assist, another part of Guru’s genAI suite of features, makes it simple for anyone to put a brain dump into a Card and have it transformed into something useful and easy to understand—in any language.
The fulfilled promise of no search
We’re at the end of an era—an era that rewarded the high effort of architecture management with endless search—and one that I don’t think we’re too sad to say goodbye to. After all, who would trade instant answers for scrolling and hoping? (🎶It’s me, hi, et cetera.)
The no search era has finally, finally arrived, and, as we predicted, it’s powered by AI. So let Guru’s AI do the hardest work for you, from knowledge creation all the way through to serving you exactly what you’re looking for, cross-referenced and ready to use.