Meet the men who help machines learn legal and financial lingo

Sajeev Aravindan and Krishna Tirupati's firm Vaultedge has developed an AI-based software platform that does what seasoned lawyers and financial experts do, but at a much faster pace

Many in the corporate world would have gone through the pain of making an important presentation all over again, after having misplaced or deleted the original accidentally. Few, however, would have spotted a business idea in the angst. But this is exactly what Sajeev Aravindan did while he was working as a global architect for storage major NetApp for backup products.

The Stanford Ignite alumnus, who also holds a Master’s degree in Computer Science, was struck by the idea of developing an enterprise search solution in 2014 after he encountered the problem of having to redo a presentation because he lost the original. After some initial brainstorming, Aravindan started a venture called Vaultedge in 2015.

A year later, he was joined by Murali Krishna Tirupati, an engineering graduate from NIT Trichy, who also holds an MBA from IIM-Ahmedabad. Tirupati was working with transportation start-up YoRide at the time, but it wasn’t long before he decided to join Aravindan and the two ended up discussing the idea at a start-up event.

Starting off with offering search solutions to enterprises, the duo quickly realised that though the solution was unique, it was proving to be a difficult sales proposition. In the meantime, both founders identified another problem statement to solve, which related to extraction of relevant information from a legal or financial document.

With the experience of dealing with enterprises, both founders discovered that lawyers and credit department executives of banks and non-banking financial companies (NBFCs) were spending hours together in extracting relevant information from a bunch of documents.

So, in early 2017, Vaultedge came up with an artificial intelligence-enabled software that not only extracts relevant information for a client, but also interprets it in different ways to ease the decision-making process. This pivoting of the business model enabled the firm to win many clients in the business process management (BPM) services and banking space.

"A legal agreement such as a contract is nothing but a bunch of clauses relating to parties, term, confidentiality or liability. So, we have developed our software with these processes by feeding various kinds of clauses in different kinds of contracts," says Aravindan. "Now, our software can interpret various legal contracts and extract the relevant information for the customer within one minute.”

Vaultedge primarily uses machine learning in this software as a service (SaaS)-based solution. which helps in extracting data points from legal or financial documents. So, when a document is processed, the software pulls out the data from documents as the first step, which then creates algorithm out of it based on coding and inference is provided to the clients.

For feeding a multitude of data points to the software relating to financial and legal contracts, Vaultedge takes the help of domain experts such as lawyers and financial analysts. It is essentially using lawyers and finance experts to help train its AI engine, which, in turn, reduces the burden on others.

Apart from analysing legal contracts, Vaultedge also provides a financial statement analysis product for banks and financial institutions. This helps in automating the whole loan processing application of any small and medium enterprise (SME) by analysing its credit-worthiness.

"Through our APIs, we give the client access to GST and income tax portals, so that they can directly download the customer's information relating to GST returns and income tax filing," says Tirupati.

"After getting the returns, our AI-based platform read the relevant information from these financial documents and interpret it to arrive at the credit worthiness of any enterprise," he added.

The company claims, the processing time of a loan is reduced to 30 minutes from an average of 1-2 weeks as usually seen in case of the financial institutions.

With such unique applications, Vaultedge is fast emerging as a preferred partner for BPM services and financial services institutions. With marquee clients like Accenture, Star and PwC, the company, Vaultedge is looking at enhancing its product portfolio going ahead.

"We currently charge our clients based on the number of documents processed and are operationally at break-even. Going ahead, the road map is to enter various other critical domains in the legal and financial services space," Aravindan added