500 million tweets, 300 billion emails or 65 billion WhatsApp messages sent every day worldwide – the information age is driving communication and the generation of data in all areas of life. Whether at work or in our private lives, every day we create a multitude of documents and data, which in turn contain information and need to be classified by us. To cope with this flood of constantly new contributions, to categorize, evaluate and file them, is probably the greatest challenge of our days.
In many professions, having the right information, at the right time, with the right person is a key factor that can make the difference between success and failure. Companies are also often faced with the problem of too much and often unorganized data from different sources. Be it e-mails, documents on a file server or data from an ERP system. The quantity increases noticeably and becomes visibly more unclear. Modern enterprise searches such as INTERGATOR solve this challenge by providing all the necessary information in one central location and freeing the user from the burden of laborious research in scattered systems.
However, the right search needs to be learned and although we now use Internet searches on a daily basis and use them as a matter of course, here too we repeatedly come up against limits that provide us with only imprecise hits. Inaccurate search queries or misleading terms deliver unsatisfactory results and leave the user frustrated or at a loss. Overly general searches result in an almost unmanageable amount of hits and no one bothers to turn to page two in online searches anymore. A common joke is “you could hide a crime on the second page of a search engine and no one would notice” – meaning: we expect the best hits among the first entries and are no longer interested in the rest.
The needle in the digital haystack
Search engine manufacturers have gone to great lengths in recent years to categorize and catalog data and documents. This often involved maintaining extensive thesauri and synonym lists, bringing a company’s language usage in line with reality. Companies often have internal language terms for products or projects that often have different meanings in the world outside the company. In addition, these exact searches with exact search terms could only return results if the term was contained in a file or document. If they did not contain this term, but were still relevant, they did not appear in the hit list.
With the emergence of machine learning as a method of artificial intelligence, a new field opened up, since for the first time sufficient computing power was available to comprehend data in its entirety. Information is related to each other by mathematical operations and the occurrence of specific terms is initially relegated to the background. This approach opens up new possibilities for a guided search in which the AI recognizes the user’s intention and selects relevant data based on this. Skeptics may be unsettled by this approach, because how is a machine supposed to know what I mean or which result is correct. Too much unknown happens in this black-box AI, one often hears as an argument here. But the search does not work in isolation and the key to successful use lies in a dialog and interaction between man and machine.
Guided search and context map
INTERGATOR SMART SEARCH extends the functionality of INTERGATOR Enterprise Search with a completely new operating concept. On the one hand, it is still possible to use filters and facets to narrow down results with just a few clicks. Files and documents are thus efficiently narrowed down according to file extension, time periods or storage locations and quickly broken down to a few relevant hits. On the other hand – and this is unique so far – search hits are conceptually transferred to a map, which locates and clusters contexts.
By selecting additional, suitable keywords, the search context is sharpened and results are refined. Starting from a rather diffuse search idea (“show me everything about topic XY”), the map visualizes related information (“XY is part of category Z”, “XY is similar to terms AB”). From here, the user can navigate step-by-step through the topic and find much more relevant hits than with a search based purely on exact terms.
The AI is therefore not a black box to be feared, inside of which some hard-to-understand magic happens, but an approach to find information beyond the narrow scope of an exact search. This does not necessarily mean that the latter is history. On the contrary, it still has its right to exist, but for those who are not always clear about which search terms lead to the right result, SMART SEARCH is a new and reliable tool for information retrieval.