Exporting Results

Exporting Results

Contract Risk Scanner for Microsoft 365 can generate structured outputs that you can download and use in Power BI, analytics platforms, or custom workflows.

What gets exported

When the agent is running with Code Interpreter enabled, a scan typically produces:

File Purpose Used By
risk_register.csv Tabular risk register data Power BI, CLM systems, analytics tools
risk_register.json Structured, machine-readable output Automations, integrations, custom tools

These files are generated during the scan and are available to download directly from the chat session.

If Code Interpreter is not enabled, the agent can still provide a text-based review but will not generate downloadable files.

CSV structure

The CSV contains one row per identified risk, with fields such as:

  • Risk identifier
  • Risk area (e.g., Liability, IP, Data)
  • Risk description
  • Severity and confidence levels
  • Supporting evidence from the contract
  • Business impact and recommended action
  • Contract name and scan date
  • Overall risk score (if calculated)

Note: Column names and ordering are designed to be consistent for analysis, but may evolve across agent versions or custom configurations.

JSON structure

The JSON output provides a structured representation of the scan, typically including:

  • Scan metadata (contract name, scan date, mode, agent version)
  • Contract metadata (if extracted)
  • Risk score (total score, band, and contributing factors)
  • Risks array (detailed entries for each identified risk)
  • Optional sections such as:
    • Missing clauses
    • Negotiation questions
    • Suggested redlines

The JSON structure is intended for downstream processing, but should be treated as a versioned output rather than a fixed contract.

Working with the outputs

In Power BI or CSV-compatible tools

  • Open or import the CSV for immediate analysis
  • Append multiple scans to build a portfolio view
  • Use the data for dashboards or reporting

In automation or integrations

  • Use the JSON output as input to:
    • Power Automate flows
    • Databases
    • Internal tools or CLM systems

Where to save outputs

The agent does not persist files after the session, so you control where results are stored. Common patterns include:

  • SharePoint libraries for team access and metadata tagging
  • OneDrive folders for individual use
  • BI dashboards or data warehouses for portfolio reporting

Versioning

Each output reflects the agent version and configuration used at the time of the scan. This is useful for:

  • Audit trails
  • Comparing results across scans
  • Understanding changes over time