Jennifer Zaino
SemanticWeb.com Contributor

Well, at least that is the goal of the vendor's latest incarnation of its semantic search engine, Cogito Answers. The technology is aimed at being a true semantic solution to addressing the challenges facing users -- both employees involved in enterprise search processes and customers trying to help themselves -- face in extracting the unstructured information they need via email, the web, and mobile devices.

"We apply the semantic engine to corporate information directly, so the system normalizes, integrates, and semantically extracts information from unstructured, where statistics say the vast majority of content or data resides," says VP of Internet and Mobile Walter Kostiuk. "The second part is the natural language interface which is definitely a breakthrough in how you extract information easily -- it reduces the technology barrier of heavy applications and lets users interact with machines as they do with people. The same way you can e-mail a colleague you now can go through the web or wireless interface to communicate with a knowledge base."

It's a tool with applicability to Fortune 500 companies in general, and to telcos in particular. Carriers have a problem in that the new age of smart phones has some downsides to all the revenue pluses it delivers in terms of increased data usage -- that is, their complexity breeds customer service calls. Lots of them. The biggest cost of selling these phones for carriers, says Kostiuk, comes in terms of dealing with customer service issues, to the tune of about $6 to $10 per call processed.

Half of those calls are just about how to use the product or service, according to research Expert has done with carriers and a smart phone maker -- and that information is already provided in a user manual and already present in companies' knowledge bases.

"So if a system existed where someone through common language could ask their questions through the web or a wireless interface, it would deflect that cost. The ROI is clear," says Kostiuk. In a beta test Expert is currently conducting with a large smart phone device maker, Cogito Answers is handling tens of thousands of questions a day, he says.
Cogito Answers takes the self-service process up a notch, he says, in a few ways, beyond providing real-time answers when it might take a customer agent 24 hours to deliver a response. Compared to keyword-search based self-service systems, which tend to deliver back lots of links -- and frustration to the user due to the irrelevancy of some of those. With Cogito Answer and its NLP capabilities, "you get the answer, not references to chapters and manuals which people get frustrated with."

Getting people the exact data they need when they need it makes them more comfortable with the self-service process, and that comfort leads them to ask more questions and get more comfortable with the product itself. That can translate to increased sales for accessories or services around the product, "but even more definable and measurable is its impact on churn," says Kostiuk. Generally users have 30 days to return a phone if they're not comfortable with using it, so it's key for sellers to make them as comfortable with the device as possible in that time frame in order to reduce returns.

True semantic engines pose challenges in terms of the fact that the richer semantic search becomes, the slower the system tends to be because there's so much more stuff to search through. But Kostiuk says the Cogito platform was built for processing speed and service, noting that it is "almost as fast as Google from a concurrent questions perspective and it is scalable just as easily by adding servers." Companies can choose to run the service behind their own firewalls, or they can opt for a software-as-a-service option.

The enterprise selects the unstructured data it wants to extend internally and externally, from a single source or multiple sources. Companies that want to make internal information available should consider dealing with incompleteness or redundancies in their information before moving to integration in order to not affect the quality of the system, advises Luca Scagliarini, vice president, strategy & business development at Expert System.

The company offers professional services as well as a toolkit to help companies in these and integration areas, but, says Kostiuk, "basically our system connects to that data in virtually any form. So there is extreme flexibility and robustness in connecting to that content."