What does the compliance lifecycle look like at your company? In globally-operating industries such as finance, there’s likely a herd of people charged with monitoring rules and regulations across countries, drafting policies and procedures for individual geographies or business units, and working to ensure controls are in place to prevent and detect violations. And that herd of individuals in some respects may be trying to herd cats, given how often aspects of compliance regulations change.
The situation presents the ideal use case for semantic technology, says Cambridge Semantics’ co-founder and VP of Technology and Client Services Lee Feigenbaum: There’s data to consider from a wealth of sources, from internal documents and control databases describing what is necessary to enforce policy at different areas and levels of the business and what reports are needed to ascertain compliance, to regulatory information published on governing bodies’ web sites or RSS feeds; people are working cross-organizationally within the company and in conjunction with the regulatory organizations; and the rules regularly change. At yesterday’s Demystifying Financial Services Semantics conference in New York City, it demonstrated its just-released Compliance Information Management Solution Accelerator, based on its Anzo semantic technology, to deliver information integration across multiple data sources, as well as an editor workplace where compliance officers or others managing these tasks can contribute and track content changes and workflow, and then seamlessly bring together the compliance content applicable to particular business units or geographies.
Semantic technology, he explained further during a Big Data session at the conference, holds the key to “horizontal big data issues,” where some data resides across many different sources. “Even if the total amount of data is very little, if it is scattered around a huge amount of disparate sources and you don’t know what you need, without semantic technology underlying you will have hard time puling information together, even if the total volume is small. That’s where a lot of the pain lies.” (See here for Cambridge’s Rob Gonzalez discussing the issue further.) Compliance information management, he said, exemplifies this problem, being an “extremely wide ecosystem of information and a lot of it is unstructured, and some of it is highly structured. Unless you can bring that together quickly and fit into models to answer your questions, you can’t get a full holistic view of your business,” which has implications for risk calculations and compliance.
With the new accelerator, companies can take individual policies and procedures and they become semantic content, he says, tagged for applicability. Unstructured data is part of the picture – for instance, for web sites where new rules and regulations are updated—through its Anzo Unstructured support for text analytics plug-ins, so that that information becomes first-class semantic data, too. Not only can compliance documents then be assembled that apply specifically to a business unit or geography, but it also becomes possible to discover and publish to employees only the changes that have occurred since they last viewed a manual.
Fixed quarterly reporting to a body such as the SEC brings enough challenges, never mind being able to deal with understanding ad hoc queries from regulatory or industry bodies. One of the things this technology can address is that executives can have up-to-date views into compliance status, Feigenbaum says. “Most people doing this are generating views of information on a scheduled basis, but now you always see the current state of compliance,” he says. “A lot of this is about cutting cycle times, sometimes dramatically. A lot of business processes have built-in cycles. You wait for them to come around and then act on something. If you make it totally up-to-date all the time, it lets you run the business a lot more proactively.”
In addition to announcing its compliance solution, Cambridge yesterday launched a new web site called Semantic University. It aims to be the education spot for semantic technology, taking a vendor-agnostic approach to getting semantic web wanna-bes up to speed. Currently the content is focused on explanations of various aspects of semantics, in both text and video, and with a more technical slant. But he hopes to add more interactive capabilities so that learners can try out some semantic tech, as well as produce content about using semantics from the business perspective.