E42 (Light Infosys) - 2020
Client
A subsidiary of a multinational conglomerate that has a presence in over 30 countries. This subsidiary is into the manufacturing of aluminium and copper, they have over 50 units in 10+ countries.
This company has numerous manufacturing plants all over India. Each plant receives a lot of visitors who come to deliver the raw materials. These goods are transported by trucks. The truck drivers have to show an invoice which is verified by security personnel. This document is later scanned and these values as manually entered into the Database. Once the security personnel has verified the invoice the truck driver's entry is allowed.With the V9 release, we enabled the users to collaborate over the APIs being developed using a Git integration. While this works fine for many use cases, the UI and functionality were limiting. - the users could just edit schema and documentation, and test in it.
Problems
In the current scenario, there is a use of manpower to verify the documents.
This, in turn, makes it time-consuming to verify each invoice.
The process of entering the details from each scanned invoice into a digital database is also long and tedious.
There is also the scope of human error
Goal
Our goal was to design a self-service booth that would automate the invoice verification process as well as provide a means to store the invoice details in the database with less human effort.
Proposed Solution
To achieve our goal we will be using OCR Scanning technology to automate the entire process.
OCR Scanning
OCR or Optical character recognition is the conversion of images of typed, handwritten or printed text into machine-readable text, whether from a scanned document, a photo of a document, a scene photo or from subtitle text superimposed on an image.
We will use this technology to scan the invoice and extract details, which be validated on the backend to confirm the entry of the truck.
These extracted details will in turn be stored in the database in a structured format.
Workflow
In the present scenario, the driver carries more than one invoice. These invoices might belong to multiple vendors. For the system to grant the user's entry, we need one verified invoice.
Execution
After speaking with our clients to understand more about our users, we decided to go ahead with a conversational approach. This way, the driver can be instructed at each step of the entry verification process.
This was a prototype of the solution, which I had created using Adobe XD.
Final Takeways
Due to India's large service-based industry, self-service solutions are not widely implemented. Introducing something like this to a group of people who are unlikely to have any experience with it will require a lot of iterations on the initial solution proposed. Currently, the solution does require human intervention, but 90% of the human effort is automated.