Sorting through endless legal documents can be a time-consuming and burdensome process, but one startup says it doesn’t have to be that way.
Semeris strives to demystify legal documentation by using the latest artificial intelligence and natural language processing techniques. Semeris’s goal is to put the information its customers need at their fingertips when and where they need it.
Semeris aims to bring structure to capital market legal documents, while providing a first-class service to customers and blending together the disciplines of finance, law, natural language processing, and artificial intelligence.
In this edition of Built with MongoDB, we talk with Semeris about how they use MongoDB Atlas Search to help customers analyze documents and extract data as quickly as possible.
In this video, Peter Jasko explains how MongoDB Atlas's fully managed service and support has been a key factor in helping Semeris scale.
Built with MongoDB: Can you tell us about Semeris?
Peter Jasko: We help our investor banking and lawyer clients analyze legal documentation. We help them extract information from the documentation that they look at. A typical transaction might have 500 to 1,000 pages of documentation, and we help them to analyze that really quickly and pull out the key information that they need to be able to review that documentation within a couple hours, rather than the 7 or 8 hours it would normally take.
Built with MongoDB: What is the value of data in your space?
Peter: Data is essential in what we do because we build models around the publicly available documentation that we see. We store that data, we analyze it, we build machine learning models around it, and then we use that to analyze less seen documentation or more private documentation that our clients have internally.
Built with MongoDB: How has your partnership with QVentures helped Semeris?
Peter: Our partnership with QVentures is not just a financial one where they’ve invested some money into our firm; they’ve also helped us uncover contacts within the market. They introduced us to the MongoDB partnership that has helped us get some credits and build out our technology onto the MongoDB platform.
Built with MongoDB: What has it been like using MongoDB’s technology?
Peter: We chose MongoDB because it’s a scalable solution, and it has a strong developer following. It’s easier for us to hire tech developers who understand the technology because MongoDB has such a strong following in the community. If we have small issues with the technology, we’re very quickly able to search and find the answer to learn how we need to resolve that.
Additionally, scalability is really important to us. And, what we found is that the MongoDB platform scales both in compute and also in storage seamlessly. We get a notification that more storage is required, and we can upgrade that online and with no customer impact and no downtime. It's really, really seamless.
Another reason we chose MongoDB is that it’s cloud agnostic. We're on AWS now, but we're almost certainly at some point going to be asked from customers to look at Azure or Google. So it's really beneficial to us that MongoDB works on all the different platforms that we look at.
Built with MongoDB: What are some of the features you use within MongoDB?
Peter: We use MongoDB Atlas Search because of its ability to retrieve thousands of data points from multiple documents. We use the indexing capability there, and the key thing that we find is that our customers want to retrieve thousands of data points from multiple different documents. A lot of our customers are analysts or investment portfolio managers, and they want that information in their hands as quickly as possible.
Built with MongoDB: What is some advice you’d give to aspiring founders and CEOs?
Peter: Try lots of things and try them quickly. Try lots of little spikes, and take the ones that work well, and eventually put those into production. Really focus on what your customers want. Ultimately, we tried a lot of different ideas, some of which we thought were great. But you have to put it in front of your customers to be able to decide which ones are really worth spending time on and putting into production quality and which ones you should just let fall by the wayside as research done but not ultimately used.