Overview
The AI Clause Analysis feature enables automated identification and tagging of contract clauses and terms using custom Contract AI Models. This capability provides flexibility to extend tagging while delivering comprehensive insights into clause distribution across your document repository.
Note that when the same P# is filtered and processed again with the same Contract AI Model, the system will delete and overwrite the previous analysis results, ensuring that only the most recent analysis data is retained.
Contract AI Credits (CAC)
AI Clause Analysis operates on a consumption-based model using Contract AI Credits (CACs). Each analysis run draws from your organization's CAC balance — meaning every time a clause analysis job is initiated, credits are deducted based on the scope of documents processed.
Administrators should ensure accounts are adequately funded before users begin leveraging the feature. Before getting started, we strongly recommend taking advantage of our AI Workshops to learn best practices associated with using this capability.
Prerequisites
Users must create a Contract AI Model in AI Design Studio before initiating clause analysis operations.
How it Works
Navigate to AI Design Studio and select AI Clause Analysis to get started. Access is managed at the Profile level by your Administrator.
Step 1: Initiating Clause Analysis
Users begin by selecting their desired Contract AI Model from the available options, with Standard Models prominently displayed. Once a model is selected, users click the Run Clause Analysis button to initiate the analysis process.
Step 2: Analysis Configuration
The Run Clause Analysis button opens a configuration dialog that allows users to define the scope and parameters for the analysis operation:
- Filtered Documents — Apply tagging to a specific subset of documents.
- All Documents — Process the entire repository.
Filter Criteria: When Filtered Documents is selected, users can apply filters such as Account, Account Type, Effective Date, Contract Type, Pramata Numbers, and other available parameters to precisely define the document set.
Once configured, clicking Run Clause Analysis initiates the process and closes the dialog.
Step 3: Reviewing the CAC Estimate
Before the analysis job begins, the system displays an upfront estimate of the Contract AI Credits (CACs) required for your specific run. This estimate is based on:
- The Contract AI Model selected — including the LLM option
- The volume of documents being processed — based on the scope defined in Step 2 (filtered or all documents).
You will see both the estimated CAC requirement and your organization's available balance. If sufficient credits are available, confirm to proceed. If not, contact your Pramata Account Manager to replenish your CAC allocation before running the job.
Once you confirm and trigger the job, the estimated CAC amount is immediately deducted from your organization's balance. However, actual processing costs may differ from the initial estimate — depending on the true cost incurred, your balance will be adjusted accordingly:
- Credit back: If the actual cost is lower than the estimate, the difference is returned to your CAC balance.
- Additional debit: If the actual cost exceeds the estimate, the additional amount is deducted from your CAC balance upon completion.
Note: Estimates are intentionally conservative to help avoid unexpected credit usage, but administrators should monitor balances regularly to ensure continued availability.
Step 4: Analysis Progress Monitoring
The Clause Analysis in Progress section displays real-time status indicators tracking the analysis workflow through four sequential stages:
- Queueing Your Request
- Analyzing Documents
- Organizing Results
- Publishing Data
Each stage is visually represented with status indicators: grey (not started), yellow (in progress), and green checkmark (completed). Click the Refresh button to view the latest progress updates.
Note: Only one tagging job can run at a time for any given Contract AI Model.
Step 5: Analysis History
Upon completion, the Clause Analysis in Progress section is replaced by the Analysis table, which provides a comprehensive audit trail of all clause analysis operations. Users can:
- Track job status and review execution details including start and completion times
- View applied filters and identify who initiated each job
- Download the complete LLM output for detailed analysis