The knowledge of an organization is widely dispersed. It may be hidden in scanned
documents like contracts and invoices. It can reside in content management systems such as
SharePoint. It may be lying in storages like Google Drive or OneDrive. To bring all this knowledge
together itself is a herculean task; but it is even more difficult to make it available to
everyone in an easy to use way.
Solutions such as datalakes make it possible to collect diverse types of data objects. However, consuming the knowledge contained in these objects still remains an unsolved problem. Tools like ElasticSearch can enable text search. But more often than not, the results returned by such queries are not suitable for the users.
Take the example of Standard Operating Procedures or SOPs. Enterprises make large investments in making and maintaining SOPs. However, they rarely prove useful when needed. If an employee needs a specific detail, they have to wade through a number of lengthy documents, trying to piece together information dispersed in many pages. We can give many examples of how organizations have all the required knowledge but struggle to make it available easily.
How convenient would it be to simply ask questions and receive answers? This would mean everyone in the organization has easy access to all the knowledge they need. Queryfic is an enterprise knowledge agent designed to help organizations achieve this goal.
Queryfic comprises two halves that work together to make such an easy distribution of knowledge possible. The first half, QfLearn collects knowledge from various sources. QfRecite, the second half is responsible for distributing the knowledge through various channels such as chat.
Enterprises can deploy Queryfic in a fully secure way so as to ensure the confidentiality of their data. They can also control what knowledge will be seen by whom. This makes Queryfic ideal for enterprises.
Queryfic uses the power of AI, especially large language models to enable knowledge sharing. Cere Labs offers various customization services to make it the best fit for any organization.
Queryfic is made up of two major components, QfLearn and QfRecite. Both the components connect to QfKb, the knowledge store.
Here is a brief description of each component:
QfLearn: QfLearn is responsible for collecting knowledge. It can connect to diverse data sources such as documents, emails, databases, servers and data lakes. The information collected from these sources is converted into forms of knowledge that are suitable for sharing. QfLearn runs continuously, either mining the sources or mining the already collected knowledge for further knowledge creation.
QfKb: The knowledge base is made up of multiple storage components suitable for persistence of knowledge. The storage is optimized for the various types of recitations used.
QfRecite: This component offers distribution of knowledge through channels such as chat, email, FAQ, recommendations and learning courses. Even though chat is the most popular channel, the other channels turn out to be the right fit depending on the purpose. Recitation can take place through native UI of Queryfic or through an API. QfRecite provides user authentication and authorization through roles.
These are huge documents which everyone finds hard to read. Queryfic can be used by the teams to search for specific terms, compare them with other similar contracts, find previous responses etc.
There is a treasure trove of documents in the customer service departments. Apart from the support tickets and past solutions, they have the help manuals, readmes and even emails from the users and product teams. Qflearning all this information can lead to unprecedented improvements in the time to solution.
A manufacturing company usually possesses manuals of its equipment. During equipment maintenance or repair, it becomes critical to find the right information from these manuals. Queryfic can qflearn all the available manuals, even if they are in hard copy form. The team members responsible for the equipment can use qfrecitation from any device such as their phone.
For equity analysts working on information about companies, the annual reports and investor call transcripts are invaluable sources. They can qflearn not only the reports and transcripts but also their own specific observations. They can ask pointed questions to help them in analysis and investing,
Read from various formats like pdf, .doc, text and many more
Extract information from scanned documents
Connect to a wide range of data sources like disk drives, cloud storage, database, datalakes, DMSs, CMSs and so on
Customize knowledge creation for special purposes
Extract metadata for access right management
Make knowledge available through various channels: chat, email, FAQ, recommendations, learning courses
Chat in a conversational style
See references along with the answer
Allow only authorized users to see the knowledge they are allowed to see
Access knowledge from mobile as well as web
Deploy within your perimeter.
Rely on the Queryfic server fully tested for penetration risk
Use the secure API offered by Queryfic
Define user access rights based on metadata
Deploy on cloud as well as on-premises
Scale with the ease enabled by microservices
Integrate with a variety of data sources
Integrate with your website or app using the API
Avail the custom integration services offered by Cere Labs and its partners