06.09.2021 | Blog Cognitive Search = Enterprise Search + AI
It is quickly explained in the dictionary: The adjective cognitive comes from the Latin word cognoscere - to know, to recognize, and means "relating to thinking, understanding or knowledge". In the context of search software, it means that the user not only finds information incredibly quickly by means of the software, but can also use this information to recognize correlations. This is precisely the benefit of enterprise search for companies - that it enables employees to quickly access relevant data in order to make decisions and perform their tasks in the best possible way.
To do this, information from all data sources must be brought together - regardless of whether the data is on the company drive or in particular specialist applications, in e-mails, in connected databases or the cloud, on the intranet or the internet. We are talking about enormous amounts of data here, which continue to grow every day.
Now can the software think or understand? Well in some ways it can. Through innovative AI-based processes and sophisticated text understanding, it not only searches data, but also analyzes it. This enables relevant core information in any content as well as documents to be identified and also extracted and cross-linked for further processing.
The best way to illustrate this is with a few examples:
- Data protection-relevant information as defined by the GDPR can be located quickly and in a fail-safe way. This means that requests for information can be answered quickly or unstructured file repositories can be cleansed of personal data.
- Resumé analysis in the search for applicants: The great advantage here lies in the matching between job profile and existing resumés within seconds in order to find suitable candidates quickly and in a fail-safe way.
- Billing of services: Intelligent software can be used to automatically analyze freetext descriptions in service reports according to activities performed. These recognized activities can be compared with the service components ticked off by the service technician so that no services performed are forgotten in the billing.
- Document topics can be identified and sorted. Incoming e-mails, for example, can be automatically analyzed and routed for reply based on content.
For a search to produce the desired results, there must be a deep understanding of language in the software. It must also understand entire question sentences and find numbers, units, proper names, synonyms, etc. For example, searching for the words "boiling point" and a "temperature of 90 - 110 °C" in the same sentence will also return hits for the sentence "The boiling point of water is 212 °F" thanks to unit normalization.
In addition, intelligent search software can be used to automatically categorize documents by keywords so that information can be efficiently retrieved and linked.
I myself have been dealing with the topic of intelligent search in corporate data for 20 years now, and I still find it extremely exciting because the field is continuously evolving. When we started using artificial intelligence in our software 15 years ago, more and more application scenarios opened up that benefit companies and government agencies. For organizations and their employees, AI-based enterprise search is simply a cool, extremely efficiency-enhancing technology that helps leverage data assets effectively.