Predictive Intelligence in Idea Portal

Predictive Intelligence in Idea Portal


Did you know that the Idea Portal in ServiceNow applies Predictive Intelligence for Innovation Management? Here is how this works.

Suppose, I have a brilliant idea and convinced that my organisation will benefit from it. What do I do? I submit the idea in the Idea portal in ServiceNow. Here is what the portal looks like.


I can quickly browse through the already existing ideas, filter them by state and/or category. A quick manual search and I am convinced that my idea is unique!! So far so good.

I click Create an Idea and start entering the data. As I start entering the idea information, I notice that I get suggestions for existing similar ideas. This is enabled by ServiceNow's machine learning algorithm.


So what's the big deal and how does this help?

It lets organisations maintain a meaningful idea database by avoiding duplicates. This removes a lot of manual scanning in the ideation process. The similar ideas are also displayed when you view the details of an idea. Use this suggestion to merge similar ideas before these are moved over to the next step.


Interesting! So how is this set up?

Open the idea module record and look for the 'Similar ideas solution definition' attribute. This defines the similarity solution that's used to find the related ideas. Open the Similarity Definition page. There are a few important items to talk through at this point.

1. Word Corpus: This is used in order to mine information in single or multiple tables for similarity or clustering. In our case, the Word Corpus includes the Title and Description field from the Ideas table. This essentially is the context used by the machine learning algorithm.

2. The table and fields section: The left columns defines the data set that will used to compare. In our case this is the same, because we want to compare the data being entered against the data in the same table.

3. The training options: This is important because you want to keep your algorithm updated based on what data is being updated in the table. This makes the algorithm adjust to your organisation and the context of your business making the solution unique!


Click Update and Retrain. You would notice the training progress in the related list.


Open the ML Solution record when this is complete. Update the Similarity Score Threshold value. This is the degree of similarity between 2 records in scale of 0-100. The solution will return a record when the score is beyond the threshold.

What should the threshold value be then?

Well, it's a matter of what the training result has been. So, open the Similarity Examples link to see the result and you will have to decide for yourself. Let's say that I am convinced that the algorithm is doing a pretty good job of identifying similarity when the score is 75 or above, then I would use that as the threshold.

At this stage, you can also test the solution. When you are convinced that this is in a good shape, you are all set to activate it. Good to go!!


So, do you have an ITBM use case where you think ServiceNow's Predictive Intelligence can help? Let us know what you think!

We are also creating some cool items related to ideation. Watch this space for more details soon!