AiThority Interview Series with Abhinav Somani, CEO at Leverton

How did you start in this space? What inspired you to work in AI for Document Management?

I used to work on wall street for investment banks and private equity firms, but in high school I had learned to computer program. It was working those late nights reading through documents manually that drove me to say, “Hey there has to be a better way!” Coupling my financial acumen and business experience with our CTO’s extensive knowledge of deep learning algorithms, we were able to create LEVERTON – a platform to use AI to get important data out of corporate and legal documentation.

How difficult is integrating AI/ Deep Learning when it comes to extracting meaningful insights from a stack of documents?

It is very challenging because documents tend to all look very different from each other, yet we often want the same information out of all of them. Everyone has their own style of creating a vendor contract, invoice, or even employment agreement – but you still need to procure the same sets of information. This is the challenge that we try to solve by creating a neural network, or artificial brain, that interprets language and text the same way the human brain does. Luckily today, the processing power is not the issue as much as the level of sophistication you need from an algorithm. IF a document repeats the vendors name 100x, but you only need to pull the Vendor name once, how does the AI know where to get that information from – the beginning, middle, or end? These are the challenges we try to solve everyday.

Tell us about LEVERTON’s products?

We have 3 solutions that work off of our 1 product. The main product has an inbuilt document upload, OCR, AI based information extraction, a reviewing section to check and edit the results, and a contract analytics platform to analyze and take a deep dive into the extracted contract data. We offer this through 3 solutions: (1) LEVERTON Assisted – our out of the box models for things like NDAs, real estate leases, vendor contracts + we take care of the human review (2) LEVERTON SaaS – our out of the box models, but you take care of the review and (3) LEVERTON AIaaS – you can create your own data model using bespoke document types and train the AI yourself.

Read More: AiThority Interview Series with Dean Abbott, Co-Founder And Chief Data Scientist at SmarterHQ

How do you differentiate LEVERTON from other similar providers?

• We are the only provider that uses true deep learning and neural network models
• We have our own proprietary OCR and deep learning algorithms – we do not use any 3^rd parties
• We have the most robust out of the box data models available for various document types
• We are the only provider to offer AIaaS – a more do it yourself approach using LEVERTON as the engine
• We pull actual structured data rather than just detecting clauses

Currently Leverton’s product is only live & available for Real Estate. In which other sectors does LEVERTON plan to expand?

Actually LEVERTON is available for multiple industries and we are already operating in industries such as a Insurance, Retail, Healthcare, Oil and Gas, Banking, Manufacturing, and many more. As our customers come to us with more and more unique documentation and data problems, we are able to help them craft solutions unique to their market. We heavily believe that any industry that has large scale manual document reading and extracting problems can benefit from the LEVERTon solution.

How do you see the raging trend of including ‘AI’ in everything impacting businesses?

AI is a toolkit that can help accelerate business outcomes. While the trend is raging, implementations have been slow and many, yet to experience the ROI that was originally promised. There’s a bit too much overselling of AI right now in the market, and at LEVERTON we try to position this as more of a “human hand meets robot hand” solution to come up with practical and achievable return on investment strategies. AI will certainly transform business, but it has to be done in a pragmatic way with clear milestones and deliverables.

As a business how do you ensure that your products comply with strict legal regulations, especially in the U.S.?

Well, I would say that we are actually even more compliant because of regulations in the EU more than the US! We are fully GDPR compliant which was a massive undertaking for companies that deal with lots of data and documentation. We have an external data protection officer whose sole job it is to ensure we are complying with all applicable regulations such as GDPR or HIPPA.

Could you share an interesting use case when LEVERTON’s solution benefitted a client?

In the real estate sector, many landlords, tenants, and service providers spend as much as 8 hour to abstract a single real estate lease to pull out hundreds of nuanced data points and important financial / legal terms. LEVERTON’s solution has been implemented at some of the top real estate institutions globally where we have dramatically reduced processing times by more than 75% and reduced costs and overhead by more than 50%. These are real numbers that are achievable through a better workflow and automation technologies. More than, we provided better structured data that could be used to analyze behavior in contracts to improve outcomes and reduce liabilities.

Read More: AiThority Interview Series with Mark Qiu, Co-Founder and COO, RoboSense

What are the biggest challenges and opportunities for the Document Management Landscape in dealing with rising technology prices?

We play in a subset of both the Document Management and Contract Lifecycle Management industries – all of whom are experiencing a renaissance. With the popularity of large organizations like Box.com, Dropbox, and DocuSign, we are going to see more and more agile, cloud based companies come up with flexible pricing that aligns a customer’s usage to the value. Implementations will get shorter and shorter, and more customers will be able to utilize out of the box solutions. While prices may increase, so will the value that is delivered – therefore overall increasing the bottom line.

Which events and webinars would you suggest to our readers as being the best in grasping information on emerging technologies?

• The AI Summit
• Future Festival
• Disrupt SF
• Machine Learning Innovation Summit

What is your take on weaponization of AI?

It’s very much possible if we don’t have the right precautions in place. AI is a tool and just like any tool it can be used for both good and bad. It is important that AI vendors not only have the right data protections in place, they also need a moral code of conduct as to how their AI can be used. We need to grow a conscience as a community and help others to adopt the socially responsible things to do.

Where do you see AI/Machine learning and other smart technologies heading beyond 2025?

AI will be ubiquitous by 2025 and we will not longer talk about it as a bespoke technology. It will be as common as the cloud is now or as enterprise applications were in the 90’s/00’s. I see it permeating literally everything we do from making househould purchases, to engaging in large scale defense contracts, to educating our children.

What start-ups are you keenly following?

I don’t really follow too many startups – I have tons of friends doing amazing things – but I really try to keep my blinders on and focus on what I’m doing right now which is making an amazing platform for our customers.

Tag the one person in the industry whose answers to these questions you would love to read.

Not sure if they mean someone from the AI community or Document/Data Management space? As an AI “guru” could be Andrew Ng

Thank you, Abhinav! That was fun and hope to see you back on AiThority soon.

Read More: AiThority Interview Series with Justin Hartzman, CEO at CoinSmart