Vaultedge automates contract abstraction and administration using AI to extract data points from documents. The system includes a dataset of about 300 data points, with the flexibility to add more as needed. This automation reduces errors and redundancy, cutting resource costs by 80%, and adapts to various workflows.
The task was to design a dashboard experience for two user groups: Admins/Managers and Consultants. The automated process needed to speed up their workflows and provide necessary tools. Admins set up the team, workflow, document legibility, quality, language, folder structure, and project hygiene. Consultants review documents, abstract content, analyze data, and identify contract issues.
Vaultedge automates contract abstraction and administration using AI to extract data points from documents. The system includes a dataset of about 300 data points, with the flexibility to add more as needed. This automation reduces errors and redundancy, cutting resource costs by 80%, and adapts to various workflows.
The task was to design a dashboard experience for two user groups: Admins/Managers and Consultants. The automated process needed to speed up their workflows and provide necessary tools. Admins set up the team, workflow, document legibility, quality, language, folder structure, and project hygiene. Consultants review documents, abstract content, analyze data, and identify contract issues.
Cognitive Overload: Users need just enough information to avoid being overwhelmed by excessive data.
Hierarchy and Speed: Efficient processes depend on a clear hierarchy in the dashboard’s content and functions.
Manual Checks: AI struggles with contextual word meanings, so consultants need to add manual notes to abstracted content.
Complex Relationships: The process is a series of interconnected steps, not linear, requiring clear breakdown to reduce cognitive overload.
Step 1: Project Setup
• Establish basic project hygiene, including folder structures, dates, and document legibility.
Step 2: Organizing the Workflow
• Assign the workflow and team, and set up templates needed for abstraction.
Step 3: Data Points Abstraction
• Extract clauses, perform analysis, and verify or update information.
Step 4: Export and Documentation
• Export the complete abstracted content and filtered results.
There Are Too Many “If Then, Else” Relationships
The first challenge was to break down the process into clear steps to reduce cognitive overload. What appears linear is actually a series of nested, interconnected steps.
Understanding both the organization and the subject domain was crucial. Contract abstraction involves two key roles: Administrator/Manager and Consultant.
Admin/Manager
• Abstract documents, group similar contracts
• Report document upload issues
• Add manual summaries to abstracted content
Consultant
• Abstract documents, group similar contracts
• Report document upload issues
• Add manual summaries to abstracted content
Wireframing and Prototyping
After finalizing the user flow, it was translated into a low-fidelity paper prototype. Co-creation sessions with this wireframe further simplified the flows. Users performed tasks, and the sessions were evaluated on:
The flow was intuitive, with users easily finding actions. Some menus/items were adjusted to match industry-standard terms.
Users quickly adapted to the streamlined process, requiring minimal onboarding. Readily available options and menu actions increased efficiency.
Creating step by step processes for projects, templates and clauses helped in maintaining the hygiene of the projects. It also brought in consistency in the processes used internally within the same organisation
Making it an all inclusive platform by providing available options pertaining to even the extreme use-case scenarios.
The interface gives users the agency to view the actual text in documents in relation to the abstracted content. They have an option to re-write or leave notes to the abstracted content.
