The Case for a Semantic Software Infrastructure
During our sales meeting the other day, I was engaged in an interesting discussion between our VP of Sales and one of our new salespeople about whether semantic technologies bring more value as an element of the IT/software infrastructure of an organization, or as a means for addressing a specific business need.
While organizations typically take the route of using semantics to address a specific business need (information retrieval, customer care, social media intelligence, etc.), I think that the real value of its adoption lies in the strategic role it can play in the IT software infrastructure.
Any organization can be seen as a system dedicated to transforming an “input” (knowledge, materials, or a combination of elements) into an “output” that is then made available to a market. While unstructured information, in terms of documentation, has always been an element of the ecosystem, in the last few years its importance has grown from a supporting tool to one of the main elements in input. This is absolutely true for certain sectors (publishing, education, digital) but it’s also becoming true for more traditional sectors like manufacturing.
In addition, the unstructured information an organization has access to has grown exponentially, and especially so compared to the structured portion. If this is true, why is it that organizations have not approached this issue in the same systematic way that they have followed for the data portion?
In the same way that they massively adopted DBMS, they should have adopted systems that would be able to process these streams with the highest possible level of performance, with a software infrastructure that would enable the highest possible speed and quality. This is how Semantic Technologies should be considered by any organization.
Yes, an organization should first identify selected and critical needs that could be addressed with semantics in order to create successful use cases. However, an evaluation of the overall investment must also look at the value of the technology from the software infrastructure point of view, independently from format or from more immediate, project-specific initiatives, and instead, how to amplify the value created.
Given the flexibility required to process the different formats of information and an organization’s unique needs, the main requirement for a software infrastructure must be that it is able to best replicate the human activity required to process such content. The most effective way to process information is to read it— Semantic Technology, with its ability to read and understand at machine speed, is the only systematic approach.
Legacy technologies used for processing text (statistics, keyword, shallow linguistics) may be able to meet the minimum level of functionality for addressing certain needs, but they will never be able to meet the most important infrastructure requirement: comprehension.