Zive

Today, we're finally sharing how we have brought classic Retrieval-Augmented Generation (RAG) to the next level in order to significantly improve generative AI quality in the enterprise. 🔍 Classic Retrieval-Augmented Generation (RAG) has been a game-changer, but it's struggling to deliver the most relevant results for individual employees. Companies that have deployed RAG-based solutions have recently shown increasing frustration with the results. 🤩 We have developed an improved concept that has proven to deliver higher accuracy and more personalized results in generative AI use-cases where LLMs need to be combined with large sets of internal data. We are introducing: Contextual Retrieval-Augmented Generation – or CRAG. Read more on our blog: DE -> https://lnkd.in/ey6TezSr EN -> https://lnkd.in/ekZkb2jj

We've solved multi-language in information retrieval! 🇺🇸🇬🇧🇩🇪🇫🇷🇪🇸🍕 Imagine you ask a question in German but you get an answer based on content from multiple languages? Zive makes it possible!

Time for a new product update! Last week we worked on dealing with duplicate content in enterprise search. We were looking for the perfect Emoji to symbolize a duplicate content. We settled on this = 💩 Because duplicate content doesn't help anybody and it starts stinking sooner than later 😜 Okay, joking aside. Duplicate content is really annoying. So companies need to deal with it. We started by showing duplicate content right on the search result page. Next step: building an automation for identifying and clearing duplicate content. Any thoughts from your side?

Today, we're happy to announce a partnership between Zive and Microsoft to deliver the most advanced and secure AI-powered enterprise search to customers in Europe and the US. English -> https://lnkd.in/e85sd9BA German -> https://lnkd.in/e4vQ6G9u

79% der Mitarbeiter sind unzufrieden mit den internen Suchmöglichkeiten. Warum ist das so? Die Idee von "Enterprise Search", einer unternehmensübergreifenden Suchmaschine, gibt es schon echt lange. Bislang wurden jedoch kaum gute Ergebnisse erzielt. Hier sind die Gründe: 📝 Mangelnde Digitalisierung: Nur was digital ist, kann auch von einer Enterprise Search durchsucht werden. 🛑 Kein Zugriff auf Daten: Die eingesetzte Software in Unternehmen ist häufig veraltet und verfügt über ungenügende Schnittstellen (APIs) für eine Indexierung. 🤯 Immer mehr Tools: Der digitale Arbeitsplatz wird immer komplexer und besteht aus immer mehr unterschiedlichen Tools, die teils die gleichen Funktionen für Zusammenarbeit und Dateiablage bieten. Wo soll man also suchen? 🙋♀️ Mangelnde Personalisierung: Je länger es ein Unternehmen gibt, desto mehr Datenmüll sammelt sich. Eine reine Schlagwortsuche reicht dann nicht aus. Die Suchmaschine muss automatisch erkennen, welche Informationen für wen gerade relevant sind. Ich glaube, dass jetzt endlich der Zeitpunkt gekommen ist, an dem mit Hilfe von neuen Technologien wie KI & Cloud der Durchbruch bei Enterprise Search gelingen wird. Deshalb habe ich Zive gegründet. Was denkt ihr?

Should the "Zive Assistant" be a separate component? 💡 Zive is a platform that helps you make sense of your company knowledge. It delivers to your footsteps just what you need, when you need it. 💬 About a month ago we launched a feature for asking questions to Zive, which then triggered the "Zive Assistant" to appear on the right side of the search screen. ❓ Now, we are asking ourselves whether we should include a separate tab just for the Zive Assistant. This could be used to start a conversation with your company knowledge right away so that you can "ask yourself to the right information". What do you think?

Today we're happy to announce that Zive has received a funding by the City of Hamburg, our home town. We'll put the money to good use 🤝, delivering innovative AI-powered knowledge management and search to organizations!

We're working on the home screen that users see when they open Zive. Let us know what you would want to see!

We've added source referencing to our Zive AI Assistant 🧐. Check it out and let us know what you think!

New website 👀 Leave a comment and let us know what we can / should improve!