Kon Leong: One customer is in the financial sector – Wells Fargo. All enterprises struggle with unstructured data, and typically they come from various areas to manage the proliferation of data and the duplication of data, which impacts storage, manage data for compliance – regulations require them to keep certain data and to be able to access it and produce it on demand. The SEC certainly has strict requirements on retention and production of data on demand. These agencies all require correspondence between brokers and clients to be readily accessible. That compliance is another need.
Another is for litigation support – to be able to produce as the courts demand it. The demand of the courts keeps rising. Before, software evidence was not the same as hard evidence and now it is. It requires every logical operation to be on top of unstructured data. Litigation can have tremendous consequences, as for example in the recent Samsung vs. Apple litigation. Samsung deleted relevant unstructured data. Some say it has had a huge impact of the loss amounting to $1.2 billion to Apple. There is a great need to manage this unstructured data.
There is an additional field that popped up a few years ago, which is records management. Every business has to maintain business records, but for hundreds of years we have only managed hard copies. However, the world has moved to digital, and the record-keeping policies and practices have not kept up with digital. There is a need for a wholesale movement to understand, retain, and extend record-keeping practices to electronic information. All of this is happening now, and the needs are concurrent.
On top of that enterprises are saying, “OK, we have extended all of this effort to do all of these things – storage management, compliance, e-discovery, records management, etc. How about if we apply analytics on all of this information and actually use it for strategic advantage?” That indeed is what is happening.
SM: Are you playing the role of just data storage, or are you also supplying the heuristics of what to do with that unstructured data?
KL: We provide the entire platform to answer all of these needs. I would hesitate to use the word heuristics because that is extending and implying something beyond simple machine learning into human-type learning. I think that is more in the vicinity of artificial intelligence. No matter what the vendors may claim, I think that is still a decade away.
SM: We have covered unstructured data at length. Autonomy has been on this series. We had lots of stories about unstructured data at various points in our coverage. One of the issues with unstructured data management and organization is that there is a lot of domain-specific structure that needs to be put in to convert unstructured data into something that can be processed reasonably. The question I am asking is if you provide that domain-specific ontology.