Codelitt and Box transform unstructured documents into actionable data

Location: Miami, FL
Industry: Technology
Use Case: Converting unstructured documents into organized data leveraging Box Skills with AI


Codelitt uses technology and user centric design to solve corporate problems with start-up speed, technology, and innovation. They focus on the build side to develop scalable solutions in platforms such as Web+Mobile, AR/VR, AI/ML, Robotics and IoT for large enterprises and offer a full stack of services from idea validation/ ideation, design, and development.

Codelitt partners with Box to create new opportunities in enterprise digital content management and has developed Ada, a custom application which utilizes machine learning and Box Skills to intelligently extract actionable data from documents.


Manual processes for data retrieval/data entry, are still prevalent in many large enterprises across a variety of industries. For most, a vast amount of data and information remains unharnessed because it lives inside unstructured documents. Businesses are often faced with the challenge of extracting the data or leaving this data to be unactionable and unusable. Extracting the data requires significant time and manual effort from current staff, taking time away from other more business-critical tasks. Not extracting this data means potentially losing million of dollars (billions across industries) in unused, unsearchable, and unactionable data.


Codelitt has built Ada, which combines Box Skills with an AI platform, Leverton, to allow you to take documents (like leases, financial documents, contracts, etc) from unstructured docs to actionable data which can be intelligently leveraged in your organization.

For now, the application is just a proof of concept, but shows what Codelitt is capable of developing. It currently allows you to upload thousands of lease agreements which it will then automatically scan and extract the data into a structured format. This data (such as rent, lease expiration, right of first refusal, termination rights, usage rights, etc) is then added to the metadata of your Box files. From within Ada you can then search, filter, and define files based on this now structured data. While the POC focuses on leases, this can be applied to any content or documents across numerous industries.

Ada uses a custom Box Skill to extract information from leases and other documents into structured data

While this process already solves a massive pain point and removes an entire manual data entry process, even more exciting is what Codelitt can now do with the data. Imagine you are an insurance carrier offering dozens of products: a new contract could be added to Ada, all of the data extracted, and then depending on the data a Box Relay workflow is kicked off for your underwriters. Perhaps you’re a large residential property manager: you could add all of your leases to Ada and all of the data from a portfolio of 100 leases is easily accessible to be analyzed for an investor’s pro forma. The possibilities are endless anytime we can harness previously unusable data.

Ada users can browse leases and filter their documents based on extracted metadata

More Information

To learn more about Codelitt solutions and services, visit: