Overview

Limble uses Natural Language Processing in combination with Machine Learning to warn work requesters that another person may have already submitted a work request for the same problem. The purpose of this is to reduce the number of duplicate Work Requests that your maintenance team receives.

This article will explain how to adjust this feature’s settings.

How it works

When someone selects the submit button on a Work Request, Limble analyzes all of the existing work orders on the appropriate assets. If Limble's analysis shows that there are Work Orders that may be referring to the same issue, Limble will give them the choice of submitting the problem anyway or canceling their request.

For example, someone fills out the Work Request form describing that it is cold in the Clubhouse and then hits Submit.

Limble detects that a Work Request has already been submitted that may be describing the same problem for the asset(s). In this example, someone already wrote “freezing in the clubhouse”. Limble then displays the alert below.

The requester will now have the option to either abandon their submission or “Submit Anyway”.

Adjust Sensitivity

Under Advanced Settings > Configuration > Work Request Settings, you can adjust how many phrases need to be similar before a warning is shown to the work requester.  You can also disable the feature entirely.

A phrase is a combination of two or more words that express a single idea but do not form a complete sentence, for example, “water spill” or “machine down”. The more phrases that you select to compare, the more specific this feature will be. For example, selecting “1 Phrase” will cause Limble to flag more Work Requests but with a higher chance of them not being real duplicates, while selecting “5 Phrases” will have the opposite effect and will flag fewer submissions but with greater likelihood of warning about a legitimate duplicate request. 

Filtering Duplicates by Groups of Assets

By default, Limble will search for possible duplicate Work Requests on all of the assets at a location. You may want to define which asset(s) or group(s) of assets will be searched for possible duplicates to increase the accuracy of this feature. This section will teach you how to do that.

Example: Building A and Building B Grouped Separately

Let’s use this two-building facility as an example:

Let’s say that you only want Work Requests in Building A to be compared with other Work Requests with Building A and Work Requests from Building B to be compared with Work Requests in Building B. This is because the work requests are in different buildings with different systems that control the AC. Therefore you will want to create two comparison groups as seen in the image below.

To create these groups, you will need to go to the settings of the asset called Building A and the asset called Building B.

Then we would need to check the box “Only check this asset (or its children) for similar Work Requests”. This needs to be done for both Building A and Building B.

Other Examples

Now that you have seen how to create two groups, one for all of the assets in Building A and one for all of the assets in Building B, let’s look at two other scenarios.

Each floor is a group - only open work requests for the floor and the two rooms will be checked for possible duplicate requests. To accomplish this, check the box labeled “Only check this asset (or its children) for similar Work Requests” on the asset titled Floor 1 and Floor 2. Because Building A has not been checked, then all assets will be checked for similar work requests. 

Each room is a group - only open work requests for the same room will be checked. To accomplish this, check the box labeled “Only check this asset (or its children) for similar Work Requests” on all of the Room assets.

In this scenario, if a Work Request was submitted for Floor 1, it would be compared to all of the Assets in this Location since Floor 1 does not belong to a group of assets.

Did this answer your question?