public

Simplifying lives of accountants with our new reconciliation engine

Before we deep-dive into our new reconciliation engine (Mentat) and how it helped improve overall process efficiency, let us understand about reconciliation in brief. What is Reconciliation?In accounting, reconciliation

2 months ago

Latest Post Addressing the Great Indian Credit Gap by Akshay Nagpal public

Before we deep-dive into our new reconciliation engine (Mentat) and how it helped improve overall process efficiency, let us understand about reconciliation in brief.

What is Reconciliation?

In accounting, reconciliation is the process of ensuring that two sets of records, usually the balances of two account, are in agreement. Reconciliation is used to ensure that the money leaving an account matches the actual money spent. This is done by making sure that the balances match at the end of a particular accounting period.

At ClearTax, we used to have a reconciliation engine for GST return filing. It was supporting only one type of reconciliation — GSTR-2A vs Purchase Register.

Challenges in our old reconciliation engine

To tackle the problems listed above, we needed a new reconciliation engine which can be customisable and would be capable to improve performance. Hence, we built Mentat — our new reconciliation engine.

How does Mentat help us?

Example,

{
"description": "Rule Description",
"task": {
"task1": {
"field_names": [
"field_name_to_be_transformed"
],
"function": "transformation_function"
},
"task2": {
"field_names": [
"field_name_to_be_transformed"
],
"function": "transformation_function2"
}
}
}

Note: We have reduced time taken to roll-out new Recon use-case from 3~4 months to 1 Week now with Mentat.

Achieving 10x improvements with Mentat

Below is the comparison between old reconciliation vs Mentat. We could clearly see around 10x improvements over old service.

Introduction to k8 Pods

How do k8 pods make a difference?

Think about it like creating server on-demand. What if we launch new pods and simply call a server for each request? We can make our reconciliation faster and taking actions, filtering/sorting data would call an independent isolated server, which would definitely be faster. K8 pods are nothing but a list of containers where each container would be running different sets of reconciliation independently.

Cost impacts of using k8 pods

Since these containers are on demand we can significantly save on infra cost. While we compare our old recon vs Mentat we could get upto 60% of cost saving.

We will talk about the Mentat architecture, server-side matching vs client-side matching, web-workers and more details about k8 pods in the next part of this blog.

Abhinav Kumar

Published 2 months ago