One of the riskiest judgement calls humans are notorious for ‘not’ making is reading through legal documents before signing or accepting them. SpotDraft, a Bangalore-based start-up, is leveraging AI to automate and streamline this lengthy contract lifecycle. SpotDraft was founded by a lawyer, Shashank Bijapur and two engineers, Madhav Bhagat and Rohith Salim. The start-up is catered towards the legal needs of individuals from different backgrounds; sales, legal and business teams, assisting them in personalised ways. Analytics India Magazine got in touch with Madhav Bhagat, Co-Founder and CTO, to learn more about the company’s unique offerings.
AIM: How is SpotDraft overcoming the problem of lengthy legal documents?
Sign up for your weekly dose of what's up in emerging technology.
Madhav: SpotDraft has multiple features that extract key information from the contract so that users don’t have to read the entire document each time they want to know something specific. In our contract summary view, the user can see the Party, Counter-party and Signatory information front and centre. In addition to such basic information, we also use AI to extract key pointers specific to a particular contract type and based on each organisation’s needs. For example, we extract Compensation, Joining Date, Probation Period, etc., for employment agreements.
Our smart templates allow business teams to create contracts that are pre-approved by Legal without any risk of stray edits and stale versions. The built-in review flow ensures that edited versions coming in from the Counter-party are always marked in smart black-line, so the lawyers know exactly what to review and don’t have to read through the whole document. Similarly, for documents that need approval, we are able to show the approver exactly what they need to approve. These features ensure that even long documents with many back and forths have the entire audit and version history stored in one place.
Further, to reduce the mundane work that lawyers have to do, we also detect definitions and errors in the document. We are able to surface any definitions mentioned in the document where it is referenced. This allows lawyers to know what a defined term means without having to scroll back through the document. The error service takes this a step further and identifies undefined capitalised or badly referenced terms.
AIM: Please explain the tech stack behind your AI platform.
We use Python and TypeScript extensively at SpotDraft – the main backend and frontend are written in those languages, respectively. The AI features rely on a myriad of technologies and languages. We use Java to parse, edit and annotate PDF files. We have a .net service that does the same for DOCX files. Once we get the text and positioning information from the files using these two services, we run models written in Tensorflow, PyTorch and JAX. In order to be able to process documents spanning 100s of pages, we use Apache Beam to run our models in an automatically parallel way without having to build the parallelisation manually. Leveraging Beam and GPUs has given us a 15x speedup as compared to CPU and custom code.
AIM: A Legal-Technological collaboration is quite unique in the Indian market. How is SpotDraft expanding this niche in the economy?
All companies are contract companies – all money coming in and going out is through a contract. Making the process quicker, faster, better – earlier, mostly enterprises cared about it, but as of now, a lot of mid-market companies and start-ups are looking to reduce turnaround time for contracting. A contract is not just associated with a legal team; it’s required by a lot of teams. At SpotDraft, we strive to make our product easy to use while still capturing the complexities of businesses. There have been other CLM tools in the past that have been mostly used by enterprises to manage their contracts. With AI and great UX, we are able to deliver enterprise-grade features even to start-ups and mid-market companies. Growing legal tech from an enterprise need – to something that can save time and money for companies of all sizes – is how we are expanding this niche in the economy.
AIM: How do SpotDraft’s services compare to a human lawyer?
SpotDraft is able to take the repetitive and mundane tasks away from the human lawyers to allow them to focus on more strategic and high-level tasks. Before SpotDraft was onboarded at a Fintech client, each NDA and service agreement signed had to be manually created by a lawyer. Using SpotDraft’s templating engine, we were able to create templates for the NDAs and service agreements so that the Sales team can create them without any legal intervention. These templates work across three entities in the US, India and the UK and only output pre-approved language for each Geography and Jurisdiction. This reduced time from 2-3 days on average to mere seconds for the first draft. This not only freed up time for the lawyers but it also reduced the scope for errors and eliminated manual copy-paste work.
AIM: How does SpotDraft work in sync with the legal sector in India?
Every Rupee that comes in or goes out of a company is due to a piece of paper signed by the company. The legal sector and the lawyers are the ones who make this happen. When architects have AutoCAD, designers have Photoshop, why should lawyers have only Word and not something built specifically for them? That is where SpotDraft comes in, and makes that process of signing, managing, reviewing, approving and tracking contracts quicker and better.
AIM: Are you planning to expand your services abroad; if so, how do you customise legal solutions to different countries?
We are already live in the US, UK, Canada, Singapore, and other English speaking countries across the world. SpotDraft can be configured as per the client’s policies and templates, so we are able to onboard different Jurisdictions and Geographies with minimal work. Our templates support input and output in multiple languages and support customary formatting for dates, numbers and currency values. We even have data centres in major regions, which ensures that customer data is stored within the boundaries of their preferred country or continent.
AIM: Tell us how the SpotDraft workspace/ team also reflects the dynamism that comes from the fact that an engineer and a lawyer founded the company.
From the early days when we used to play Lawyers vs Engineers games in Call of Duty and at the bar, we have always tried to embrace the differing backgrounds and get people to learn from each other.
Now SpotDraft lawyers are able to code up templates in Word, so we fondly call them legal engineers. And on the engineering side, we have seen a lot of AI team members being able to not only understand contracts but also extract rights and obligations from them as they have learned the ins and outs of contracts in their attempts to train the AI to do the same thing.
SpotDraft also has people who have crossed over from being lawyers to being product managers as they are able to bring their expertise related to user needs more naturally.
AIM: How do you see the Indian legal space evolving with AI in the future? What are SpotDraft’s expansion plans in this context?
There have been experiments in using AI across the legal domain. Researchers have applied it to the judicial system, which is particularly useful in the Indian context due to the huge backlog of cases in our courts. I believe we can apply AI to make law more democratised; give everyone more equitable access to legal tools and information. We are currently focused on solving this for the companies with regards to their contracts, like analysing contracts against their playbook and showing them the good, bad and ugly, but would love to apply the same techniques and solutions to individuals and non-profit entities.