INFORMATION TECHNOLOGY

Interview: IBM's Marc Andrews on UIMA, OmniFind, BI, and Search

July 12, 2006

In the past six months, Cognos and IBM have made several significant announcements in terms of a new enhanced level of partnership and around new functionality for enterprise search. Cognos Senior Writer Kelsey Howarth caught up with Marc Andrews, Program Director, Strategy and Evangelism for IBM Information Management, for his take on how these developments will help joint customers.

The power of partnership

KH: In March 2006, Cognos and IBM announced a new strategic alliance. Can you tell us a bit more about it?

MA: "The new Global Strategic Alliance extends and expands our existing partnership. Cognos is the first business intelligence vendor to achieve this level of partnership. At IBM we're aware that a lot of companies are generating extreme business value from integrating Cognos technology with various different IBM technologies, services, and hardware.

As a result of that we made a decision to further invest in this partnership to increase the level of integration between our two companies and work as closely as possible to develop and deliver additional solutions that drive value for our customers."

The search for meaning

KH: In April 2006 Cognos and IBM announced integration between Cognos 8 Business Intelligence and IBM WebSphere Information Integrator OmniFind Edition. Why has IBM chosen to partner with and integrate with Cognos for enterprise search?

MA: "We found that organizations were investing a lot into managing and capturing all of their business intelligence information but they had separate initiatives to enable users to search for business critical information across the organization. Since the two processes were not aligned and integrated people were not able to access all of that business intelligence data as part of their enterprise search initiatives.

Cognos and IBM knew that by combining enterprise search technologies with business intelligence could provide a more simplified way to make that knowledge available to broader audience. Cognos has been at the forefront of enabling advanced search capabilities of the business intelligence data so we worked together to integrate OmniFind and Cognos 8 BI for content discovery.

This allows customers to do a single search to find all enterprise knowledge, whether it's locked in corporate intranets, Web sites, enterprise content management systems, or other content sources, or within their BI environment."

The value of OmniFind

KH: Can you give us a quick overview of OmniFind and its value to customers?

MA: "OmniFind offers robust content integration capabilities, but it extends to extract information from that underlying content so that you can identify concepts and facts and other critical knowledge that isn't always accessible through simple keyword search.

On top of that, we added another layer to provide the information within context and enable users to navigate the information that they find. What we found was that the standard list of links or results that is typically provided by most search environments is not enough for users anymore. They need to be able to navigate their search results and they want the search results and have them delivered in the context to the other applications and processes they are working with.

OmniFind provides all these capabilities together to deliver more holistic enterprise search capabilities to pull in all enterprise content to extract content buried within that content and enable more powerful information access capabilities."

Bringing it all together

KH: What is IBM's position on convergence of structured and unstructured data for BI and enterprise search?

"This an interesting topic because in the past we've seen that BI has traditionally focused only on structured data. As a result it has not accessed much of the unstructured information and knowledge buried across the organization. This can include correspondence, invoices, contracts, and design documents, but also be free-form text fields within the structured data sources like customer comments, technician comments, or the description or notes fields in a customer call centre application.

By applying some unstructured analytics capabilities we can extract knowledge from these unstructured content sources and make this knowledge available in a structured format that can then be leverage by our more traditional business intelligence tools. This is what we are doing with Cognos.

We're enabling our customers to extract information from unstructured content so they can deliver more advanced reporting and analysis using their existing Cognos BI tools. In doing so, they can deliver new business insights that were impossible to generate in the past."

OmniFind and UIMA

KH: In terms of search there has been considerable industry reporting around UIMA. What is UIMA and why is it important?

"An unstructured information management architecture, or UIMA, provides the first open framework for plugging in different analytics capabilities to extract knowledge for unstructured information. UIMA makes it easier for organizations to combine all of these different technologies to deliver more powerful business solutions.

It also touches on the last point we were discussing. That is the fact that traditional search has only looked for keywords within dimensions of words and a phrase within documents, but it has been unable to extract facts and concepts from the underlying unstructured information.

UIMA allows for people to not only search for mentions of people, places, organizations, products, parts, specific problems and concerns, but also allows for insight into relationship between elements and the context in which they are used. As an example, was a person located in a place or just talking about a place? Did a specific condition cause a problem or did it prevent the problem? This is the type of knowledge that needs to be extracted from unstructured information. This is the power of UIMA and it is the framework that IBM OmniFind employs."

BI on Demand

KH: How does Cognos help IBM deliver on its vision for Information on Demand?

MA: "Information on Demand is all about making the right information available to the right people at the right time. And it's about delivering that information within the appropriate business context and in line of various different business applications.

As part of that what companies are looking for is the ability to do analysis of those business activities at the front of the process as opposed to at the end of the process. This is what is enabling organizations to transform their business processes and use those processes as competitive advantage.

When start treating their information as a real asset they can create competitive differentiators and drive significantly more return on their IT investment. What Cognos is doing is providing those analysis capabilities on the top of the structured and unstructured information that IBM is servicing out of these back-end systems.

While IBM makes it easier for organizations to extract information from various different underlying sources and make it available to various different business processes, Cognos is enabling customers to do more advanced reporting and analysis of that information across the organization."

About Marc Andrews

Marc Andrews is Program Director, Strategy and Evangelism for IBM Information Management In this role Marc is responsible for driving strategy and business development for the IBM Content Discovery portfolio, which includes IBM products for content integration, text analytics and enterprise search. In the past Marc has also been for developing the partner ecosystem around UIMA (Unstructured Information Management Architecture), an open framework designed to provide interoperability of various text analytics components. Visit Marc's blog for more information on IBM, UIMA, and more.