The APQC Blog

How Knowledge Management Knows If You Are Interesting

The technical answer to “What is interesting?” was interesting enough that IBM took out a patent on it in 2013. It should also be of great interest to you…that is, if you want to get recommended for a project, lure an executive to open an email from you, protect yourself from unwanted email, find friends and people with similar interests, or generally get noticed in a largely anonymous and virtual organization.

Let me explain.

We recently finished our 11th Knowledge Management Advanced Working Group (KM AWG). This one focused on breakthroughs in expertise location and recommending high value content to employees. Our goal was to explore how advanced computing and machine learning could give custom tailored – ”bespoke”—answers to someone’s search questions regardless of whether they had a need for expertise, or a great article or the answer to a question.

IBM took this concept of custom responses one step further. Using the digital breadcrumbs left behind as you work and interact with others, their cognitive computing applications can run predictive analytics calculations to serve up what should be “interesting” to you and save you from the boring stuff you typically ignore anyway. And of course, your colleagues can protect their precious mindshare from your spam and “reply-alls.’

For this to work, a software designer and programmer has to code and weight variables with ‘If-then” statements. Based on data about you, the computer can run a mathematical calculation of “interestingness” of something to you. Variables that would affect your interestingness index might include how frequently you have mentioned a topic online or have searched it, and whether you have interacted with someone, especially on that topic. Further, it might give a little weight to what others who searched on that topic ALSO searched for (e.g., potential adjacent topics). The people they interacted with on that topic are also potentially interesting. And so on through dozens of other variables.

Weighting these variables is where the magic and the machine learning come in. By giving the system “feedback” on the results (e.g., click or don’t click) the system tweaks the algorithm and weights. It “learns.”

As it learns, it saves you time and points you to people and content of interest to you.  

Louis Richardson, one of our keynoters at our upcoming 2017 Knowledge Management Conference, puts the value of “interestingness” this way:

“If someone is calling you, or asking for a meeting, or sending you an email, we can put that in context of, 'Have you paid attention to this person in the past? How do you know this person?' Are there certain themes or topics that you talked about recently that we should bring to your attention? As you open this email or meeting invitation, what [else] can we show you to help you answer those questions so that you could actually pay attention to the content of what’s there? We are trying to answer those questions for people so that when they engage, they know the answers.”

He went on to say that the system can develop a filter that ignores things. If it, the system, sees that you’re ignoring suggestions or people, then you’re not interested. Why even bring it to your attention because you’ll just ignore them again?

But, Richardson points out that interestingness is not just about keeping things out; it is also about alerting you to things YOU might be interested in or should know about. “This might be very interesting to you because of the question you just asked.” Or “This is an area you might want to take a look at to find a team member or a piece of knowledge.” (I call this “anticipatory knowledge delivery of adjacent topics.” Say that three time quickly.) 

It’s not just about saving you bits of time and distraction. It’s about what is actually making your time more useful.

Come to the APQC’s KM conference April 27-28, 2017 to meet some of the most interesting people in the world and become one of them.