Lost data? Just AskGordy

Gordon Flesch Co. partners with IBM to integrate Watson AI and natural language processing capabilities to solve a challenging business problem and allow users to search any electronic content management system.

These days, for just about anything you want to know — but don’t — the answer is invariably: Google it.

Unfortunately, if you’re searching for a document, file, or other piece of electronic information in your company’s enterprise content management system, Google won’t get you very far. So, instead, why not AskGordy?

That’s what Madison-based office technology solutions provider Gordon Flesch Co. is hoping everyday workers in any organization will start doing. Gordon Flesch unveiled its brand new AskGordy application July 16 to help solve one of the thorniest problems in document management — how to make the exploding volumes of data buried in corporate electronic content management platforms searchable, accessible, and more useful.

Mike Adams, development manager for Gordon Flesch, was the brains and driving force behind the project, which utilizes the power of IBM’s Watson, perhaps the world’s most notable artificial intelligence platform.

Adams says Gordon Flesch has been hearing from customers for a long time that they have invested a lot of time, money, and energy into enterprise content management systems, but that they weren’t getting the expected value out of them. “And according to research, 80% to 85% of the data in these kind of data collections is ‘dark data,’ which means it never gets found or used after it disappears into the database,” Adams notes.

“We have also been watching the development of Watson and other AI tools, and realized that we could bring the new cognitive capabilities to our customers’ historic document management solutions,” continues Adams. “I don’t know that there was any single ‘aha’ moment, but it became obvious to us that Watson’s cognitive capabilities could help organizations discover the knowledge trapped in their document collections.”

Because Gordon Flesch has enjoyed a long application integration relationship with IBM, the company had early access to some of the Watson development tools as they started to become available for use. Adams says Gordon Flesch built a preliminary pilot with those tools in 2017, and then showed IBM the concept and working model to get additional technical assistance.

“IBM got excited about the project because their own internal Watson-based applications are geared for very specific, high-investment applications, such as the Watson Oncology project,” says Adams. “Gordon Flesch was taking the same base functionality and applying it to a much wider range of users that IBM doesn’t often reach. Most of our customers are probably not big enough to collaborate with IBM on a Watson-based tool, but with us in the middle working on the development of AskGordy, we could deliver Watson’s capabilities to just about any sized organization.”

Throughout the development, IBM offered technical, marketing, and administrative assistance, says Adams, and 24 separate IBM employees are currently assigned to the AskGordy Success Team.

Combining search with intent

The public first became aware of Watson with appearances on Jeopardy! several years ago. Massive investments in time, development, and acquisitions followed, improving the natural language understanding and cognitive response capabilities daily.

AskGordy attempts to take Watson’s intelligence to another level, to the point where users help train and refine the quality of both query and responses.

“The fact is, no cognitive system is perfect out of the box,” explains Adams. “But the advantage is that it gets better and more accurate with every search. In essence, the system learns and becomes more accurate with each additional data set and user interaction.”



So, what does that mean for users? Current document storage and management systems rely on hierarchy, common search functions, and historical user knowledge as their principal retrieval mechanisms, notes Adams. In each of those cases, the response is based on some measure of matching the search term with content.

AskGordy extracts the user’s intent from the query and uses natural language understanding to determine content within the company’s document stores to provide references that meet that intent, rather than matching the actual words searched.

In so doing, the estimated 80% of knowledge stored in these documents move from being dark data, with little to no chance of retrieval, to accessible and valuable.

“For example, assume you are investigating an insurance issue and you need documents pertaining to insurance claims related to young people,” posits Adams. “If you search for terms like ‘young people,’ most ECMs will return documents with that term. Of course, no one writing about young people uses the term ‘young people.’ However, a cognitive system like Watson can find documents with related words like mother and father, and words for activities correlated with children, like baseball and football.

“It becomes even more useful if companies have stored documents in multiple systems and file shares, as most do,” Adams continues. “Rather than being limited to the dedicated search on each, AskGordy applies the same intent-based query to documents in any of those storage locations.”

Kin to Alexa

Consumers are already becoming familiar with voice-activated devices and applications such as Alexa or Google Home, notes Adams. Those are consumer oriented with a nascent — but growing — ability to infer intent in the user’s interactions.

With AskGordy, Gordon Flesch hopes to put the power of AI into the hands of the everyday worker in any organization. Best of all, AskGordy users can search their data sets using voice commands and have results returned to any platform, whether it is a computer, tablet, or mobile phone.

“It enhances, rather than replaces, existing search engines, giving users different options and ways to drill down into their datasets,” explains Adams. “In the case of AskGordy, AI can be said to stand for augmented intelligence, increasing the accuracy, reliability, and ease of use of the search technology.”

The AskGordy app interface is available for free download in iOS and Android, as well as available for download and onboarding on the IBM Marketplace.

Click here to sign up for the free IB ezine — your twice-weekly resource for local business news, analysis, voices, and the names you need to know. If you are not already a subscriber to In Business magazine, be sure to sign up for our monthly print edition here.