It hasn’t been long since the wireless industry decided to expand it’s high-speed wireless services, promising easy access to corporate data from a mobile phone (so called wireless data extensibility). The industry-wide goal is a potential factor of improvement for workforce effectiveness and efficiency in enterprises worldwide.
Progress can already be seen as handset manufacturers launch smaller and faster mobile devices with touch screens, keyboards and new operating systems for on-the-go access to content. With the infrastructure in place, one can expect end users to be easily connected to relevant content, thus fulfilling the wireless data promise. Unfortunately, the unique characteristics of mobile phones and their intelligent counterparts (smartphones) still pose a challenge when it comes to offering the ideal end user experience. For one thing, the latest evolutions in graphical user interfaces (GUI), such as those provided by Apple’s iPhone, provide decent access to menus, content lists and cursory internet search, but they really just mimic the same experience users have when accessing applications from a regular computer. Secondly, due to the small screens, people who use the text messaging features to access data from a mobile phone require a user interface that’s simple, with a built-in Q&A mechanism providing just the minimal amount (possibly the only) information requested by the user. That would be smart.
This is why I believe that well-developed, high-performance, natural language interfaces are the ideal solution to enable users to access the information required. And from our personal experience alongside a leading company in the mobile communications market, we can safely say that users seem to agree. Just in the last few months, the service we developed for this company has been flooded with more than 100K questions daily. Real semantic technology in addition to a well-designed knowledge base is a recipe for success which can reach level of performance (90% of correct single answers) that is no match for any version of a generic keyword-based system.
Author: Luca Scagliarini