Customer delight … is not enough ?

Customer delight … is not enough ?

 

Friends… Don’t get me wrong, I don’t mean to say that Customer delight is not important …All I am saying is it’s not enough to stop at that …

Every business tries to exceed Customer expectations to create a positive customer experience.

But, when you are a B2B company, the game is different. It is not enough to delight your customers, but you also have to ensure that you go beyond it and focus on “Delight your Customer’s customer”.

And the reason is – unless you solve the problem that your User – the customer’s customer is facing, you can not really make it even to the customer satisfactionlet drive the delight for your customer.

It is only if the problem of the user is solved when they are going to love your customer, bringing a delight.

At Coreview, we extend this delight down the chain, knowingly and purposefully.

One of our existing customers, an enterprising company with their own Script-less Test Automation Platform, was facing an interesting challenge.

Their customers are big insurance companies who need to process a huge number of PDFs in their workflow. For testing these critical workflows, a lot of test data comes from these PDFs. This was done manually, to begin with.

But with the increasing number of workflows to test and a huge number of complex PDFs needed for test data, it became difficult for the employees to keep up. The process became very daunting. Errors started creeping in Custom Scripts used for data extraction from PDFs, the process started becoming unreliable.

This snowballed into a lot of negativity.

The customer’s customer started being unhappy! Warning signs for our customers.

CoreView came up with a solution powered by ML.

A data pipeline was created, that converted the PDFs to HTML, processed these HTML, and used ML on these HTML for classification and extraction of various sections and their contents.

The solution also extracted data from simple tables and complex, merged tables, and presented it as programmable collection objects.

With the help of this solution, they were able to extract data from various PDF formats and further transform them into simple and easy-to-understand CSV files and easily consumable objects.

This solution allowed to graphically selection of multiple data values from a PDF and label, group it, as customized objects. Such objects were then extracted multiple similar PDFs, with a confidence score associated with each object extracted.

The User experience of such a system is very critical since the users of this system have to buy in the results, and also readily give feedback to help refine the ML models. We used to confidence scores to gamify the user experience and chide the users to give us feedback on the results.

Today, the manual efforts of converting PDF content to consumable data, for thousands of data is replaced by complete automated workflow, with more than 70% accuracy, and the remaining 30% is reported exceptions which are corrected by minimal manual interventions in the system.

Seeing the amount of manual work saved, and the increasing accuracy of the system, our customer’s customer participated very actively in the feedback process.

This is helping in refining the system, over time. This not only improved the accuracy and the speed but also delighted our customer’s customer. For our customers, we created value by delivering a huge competitive edge.

Our customer was happy because they were able to generate great value to their customers which resulted in increased business for them.

Today the system and model is being tweaked to handle more variations of PDFs and creating similar workflows for different industries supported by our customer.

For the Software services industry, it’s important to think like the customer’s customer and keep his ROI at the center of the solution.

This will automatically keep your customer delighted.

The other way round is assumed, but not always a given, Do you agree?

Share this post

Leave a Reply

Your email address will not be published. Required fields are marked *