COVID Impact - Logistics Service Providers - Change your game plan- Part II - Make Magic from your Data

Human data entry may not be sustainable in the future and businesses need to shift towards data entry automation with AI/ ML. The world is getting increasingly automated with every passing day, and the common errors due to data entry have to be eliminated.

Although elimination of manual errors is the major reason driving data entry automation, following are some of the other factors that are worth evaluating, when it comes to automating data entry :-

1. Saves Enormous Time - automating data entry saves significant amount of time, thereby slashing down the turnaround time of a project
2. Makes Data Entry Effortless - data entry automation effectively eliminates the daunting task of making manual entries, saving valuable man-hours
3. Eliminates Human Errors - being an advanced technology, automated data entry is incredibly accurate leaving no scope for manual errors
4. Keeping Up with the Industry Standards - with numerous businesses automating their data entry tasks, it is important for a business to implement data entry automation to stay on power with the competitors

How much do you know your data. There is so much gold hidden in your data. Dig and Refine ir to make it valuable. Lets understand this with some real life examples

1. A customer requests you for a pick up for transportation . You ask the customer the details of the pick up location , delivery location , the number of handling units , weight , volume , commodity details , pick up time , delivery date and time , delivery contact details and any other information that is required. This would envisage capture of 10 to 15 fields and more than 100 plus key strokes. This takes time.
2. Normally data entry is aided by drop downs. These drop downs are loaded with all the hundred of data options and the data entry operator keys in the first few keys to narrow the options. This too takes time.
3. The process above may take from 2 to 5 mins depending on the interaction and how readily data is available and its form and shape.
4. Each of these data points are inter related to each other much more than one can imagine.
5. Create a matrix of all the data points of a single screen where entries are made by putting then in row and columns. Establish an inter relationship as to how many times was the data element relation the same majority of the times , sometimes and least amount of times. This creates a heat map ( see the header image as a sample ) and and helps create the inter relation prediction capability.
6. Having established this relationship one can easily find one or two key data fields that drive the maximum inter relational predictability
7. These one or two fields become your data entry keys - in simpler terms if you enter these one or two of these key data fields the other data will get automatically populated with high predictability
8. One now deploys an AI / ML tool to execute this in the system and magic happens. You just enter one or two fields and get the other fields to be filled in the system. This saves 90% of the data entry effort.
9. Even in a few cases wherein the data fields cannot be predicted accurately the AI / ML engine brings the drop down options from the hundreds to the top three options making it effortless for the data entry operator to do the selection
10. Things become better as you keep accumulating more and more data and the AI/ ML engine learns more and reinforces the data learning

Some sample data that we worked on revealed the following

1. For transportation booking requests the high prediction capability using only one key field was 86% with only one years data being mined.
2. For Advance shipping notes data entry in WMS the high prediction capability using only one key field was 94% with only one years data being mined.
3. For Sales Orders data entry in WMS the high prediction capability using only one key field was 88% with only one years data being mined.
4. If longer time period data is mined and AI / ML learning is reinforced the prediction percentages will shoot up.
5. The savings are obvious and need no elucidation
6. The use cases can be as many as your imagination can think of...

Your data contains more gold than you know ... Mine it and not allow it to be wasted

We will make this more interesting in the next article