6 weeks
from idea to MVP
from idea to MVP
raised on total funding
San Francisco / New York
Product discovery
UX/UI design
Product management
Development
Testing
Based in California, Tome is an innovative legal tech startup with a game-changing product: an application that harnesses the power of Artificial Intelligence and Large Language Models to distill complex investment contracts into simple, helpful, and data-rich summaries.
For Tome users, it's a seamless experience—upload a contract and, within minutes, receive a clear summary that emphasizes any non-standard provisions. Should any section raise flags, rest assured—Tome’s legal team will review and clarify it for you within a few hours.
Sudolabs stepped in with a mission: empower TOME's legal team with an internal tool tailored for additional contract assessment and modification. By doing so, Sudolabs not only streamlined their workflow but also paved the way for TOME's engineering team to zero in on perfecting the machine learning model. Another example of successful synergy!
Tome was founded in 2021 by Stephen Trusheim and Nadia Dugal with the ultimate goal of democratizing access to contracts.
Stephen Trusheim, an AI engineer by background, transitioned into VC funds where he later became the Partner & Head of Operations at SignalFire. Nadia Dugal was a leader at Flourish Ventures, the Omidyar Network, 500 Startups, and she is also an educator on venture deals for Berkeley Law and the National Venture Capital Association.
The Tome ecosystem consists of microservices for contract processing, a customer-facing app for contract summaries, and various Retool apps for developers. However, the ecosystem lacked an internal tool for lawyers to view, review, and perform additional operations on contracts themselves.
The initial requirements from Stephen were very clear:
Diving into our journey with Tome, it all kicked off with a vibrant two-week product discovery sprint. Curious about what 'product discovery' is? Check our article written by our CPO.
In this phase, we closely connected with Tome's engineers, lawyers, and even the CEO to understand their workflow and expectations for the internal review tool. We took care of proper user experience and ease of use. This deep dive set the stage for the MVP.
What followed were dynamic four-week development sprints, racing to bring the MVP to life. A team consisted of a product designer, two ace developers, and a budding junior dev.
The cherry on top? This month-long endeavor blossomed into a year-long partnership, a testament to our mutual commitment and synergy.
Rapid feedback in MVP phases is essential; delays dilute value. We showcased even partially complete features, utilizing feature flags for in-progress elements. Throughout our partnership, the client had new features deployed on production WEEKLY. These regular releases and demos ensured real-time client feedback, fostering team motivation and ensuring the client felt consistent progress.
Each week, our team:
Sudolabs supported the design & development of the internal review application for Tome’s legal team - a tool that enables rapid human intervention and processing of discrepancies in the output of AI-generated contract translations.
Having the tech team actively involved in product discovery was a game-changer. The tech lead, in particular, played a key role early on, helping everyone understand the product and its tech challenges.
More meetings, including those for new features, with active developer participation, significantly improved comprehension of tasks. Demo meetings involving all developers allowed quick responses to client needs, reducing the need for extensive knowledge transfer between the product manager and the development team.
Additionally, giving engineers creative freedom by explaining the "why" behind features and allowing flexibility in tasks proved valuable for enhancing product value and driving proactive action. One example can be the implementation of a viewer for the difference between two documents. Since devs usually work with code diff tools daily and they got the freedom to enhance the product, they were able to drive the whole feature and implement improvements beyond the client's requirements.
Despite operating within a familiar tech stack, which included NextJS, Typescript, Chakra UI, Postgres, and Apollo GraphQL, it's noteworthy that the client initially utilized Python for most components. Consequently, we extended our support by making minor contributions to their Python backend. We also took the time to delve into the intricacies of the existing codebase, familiarizing ourselves with key aspects of the business logic.
In TOME’s case, we give precedence to process optimization rather than code optimization when aiming for swift product delivery. Key components include automated deployment and checks within the CI/CD pipeline. We employ tools like ESLint to automate the resolution of recurring code review issues, facilitating faster development and enhancing overall code quality.
We tweaked the design to match our existing Chakra UI library, as we were all about fast delivery. The designer wasn't always thrilled, but we can always fine-tune certain UI elements as needed. For example, we adapted tooltips and badges to work with Chakra UI, even if they were originally designed differently.
Next.js (React.js)
Chakra UI
Apollo GraphQL Client
Chromatic
PostgreSQL
Apollo GraphQL Server
Jest
DataDog
44 Montgomery St, San Francisco, CA 94104
