As Web 2.0 applications like price comparison sites, travel planner sites, mashups become more commonplace, consumers are becoming used to and expecting features of the semantic web in their everyday lives – albeit without realizing their presence.

These sites exist because of the interweaving of existing technologies - federated search or content surfacing engines and structured data and/or metadata over unstructured content.

The rumour circulating last month that Google is to buy ITA Software [1], a firm who provide airfare pricing, passenger booking and reservation management systems, is a good example of the power that can be derived from owning the structured version of what is generally available in an unstructured form.

Flight information is an already exceptionally well formatted information asset when compared with the more traditional amorphous silos of unstructured internal enterprise content.

Employees are coming to expect the same level of functionality from enterprise systems as they are getting from-public facing internet systems so there is internal pressure to invest. This degree of information classification and retrieval is driving enterprise adoption of semantic technologies.

With growing pressure to produce semantically annotated information, enterprises will have to invest to ensure their information production systems are capable of creating and carrying metadata for their otherwise unstructured information. But the infrastructure investment is only one side of the effort - without corresponding investments in information architecture, ontologies, classification strategies, etc. the technology will not deliver the expected benefits.

What else should you consider? You need ontologies, classification, metadata management, implementation of micro-format standards like RDFs to get to a base level. Ideally, encompassing all of that is an Information Architecture (IA) strategy that is forward thinking and points to public descriptors so that by the time crawlers, agents or similar consumers of data read your content they can position it in the wider context of the information they have collected from other sites.

All of that isn’t out of reach, but it is still hard work to piece together, and in most organisations, still very difficult to justify for internal-only systems.

Hopefully enterprises will realise soon enough that classifying and annotating information at source is far more economical than post-tagging it for public consumption: post-tagging requires additional, complex, dedicated business processes if sufficient quality controls are implemented.

Embracing semantic technologies and adopting a good strategy to implement them should drive many internal productivity benefits derived from the structuring and subsequent improved findability and re-use of the information and see increasing enterprise adoption of these technologies.

[1] http://www.reuters.com/article/idUSTRE63K0MT20100421 accessed 25 May 2010