Contract Portfolio Intelligence

Contract Portfolio Intelligence

Performs portfolio-scale contract analysis: batch extraction of terms, obligations, and risks across a large set of agreements, plus playbook-based redlining against organizational standards.

Prerequisites

Before starting, gather:

1. Contract corpus — uploaded documents (PDF, DOCX) or a vault/folder reference containing the agreements
2. Analysis scope — what to extract (renewals, indemnities, payment terms, liability caps, termination rights, or custom fields)
3. Playbook (if redlining) — the organization's preferred positions, fallback clauses, and non-negotiable terms
4. Output format — table/spreadsheet, narrative report, or both
5. Audience — who receives the output (GC, board, finance, procurement)

Workflow

Phase 1: Ingest and Classify

1. Process each contract through OCR if scanned; extract full text from native PDFs and DOCX files
2. Identify contract type (NDA, MSA, SaaS agreement, employment, vendor, lease, SOW, amendment, etc.)
3. Detect parent-child relationships (master agreements → amendments → SOWs → change orders)
4. De-duplicate — flag identical or near-identical documents
5. Tag metadata: parties, effective date, governing law, contract type

Output a contract inventory table:

Phase 2: Term Extraction

For each contract, extract the following structured fields (skip fields that don't apply to the contract type):

Dates & Term

• Effective date, expiration date, term length
• Auto-renewal (yes/no), renewal notice period, notice deadline
• Termination for convenience (yes/no), notice period, cure period

Financial Terms

• Contract value / total commitment
• Payment terms (Net 30, Net 60, etc.)
• Price escalation / adjustment mechanisms
• Minimum commitments or volume thresholds
• Late payment penalties

Risk Provisions

• Indemnification — mutual vs. one-sided, carve-outs, cap
• Limitation of liability — aggregate cap, per-incident cap, excluded damages
• Insurance requirements — types, minimum coverage amounts
• Warranty — scope, duration, remedy limitations
• Governing law and dispute resolution (litigation vs. arbitration, venue)

Obligations

• Deliverables and milestones with deadlines
• Reporting or audit rights
• Data protection / security obligations
• Non-compete or exclusivity restrictions
• Confidentiality term and surviving obligations

IP & Data

• IP ownership and license-back provisions
• Data handling, return, and destruction requirements
• Usage rights post-termination

Store extracted data in a structured format (one row per contract, one column per field).

Phase 3: Portfolio Reports

Generate the reports requested by the user. Common report types:

Renewal Calendar

• All contracts with renewal/expiration dates within a specified window
• Sorted by urgency (nearest deadline first)
• Flag auto-renewals where notice deadline is approaching
• Include counterparty, contract value, and required action

Risk Matrix

• Each contract scored on risk dimensions: indemnification exposure, liability cap adequacy, insurance gaps, unfavorable termination terms
• Risk level per dimension: High / Medium / Low
• Aggregate portfolio risk summary
• Prioritized list of contracts requiring renegotiation

Obligation Tracker

• Active obligations by contract with deadlines
• Grouped by responsible team (legal, finance, ops, IT)
• Status: upcoming / due / overdue / completed
• Escalation flags for overdue items

Financial Summary

• Total portfolio commitment by period (quarterly/annual)
• Breakdown by contract type, counterparty, or business unit
• Contracts approaching minimum commitment thresholds
• Payment term distribution

Custom Extraction

• User specifies any data point via natural language prompt
• Agent extracts that field across all contracts
• Returns structured table with contract reference, extracted value, and confidence

Phase 4: Playbook Redlining

When the user provides a playbook of preferred positions:

1. Parse the playbook — extract each preferred clause, its fallback positions, and non-negotiable terms
2. Compare each contract against the playbook provision-by-provision:
   • ✅ Matches preferred position → no action
   • ⚠️ Deviates but within fallback range → flag with suggested language
   • ❌ Violates non-negotiable → flag as critical, provide replacement language
3. Generate a redline summary per contract:
   • Section reference, current language, preferred language, risk level, recommended action
4. Produce a portfolio deviation report:
   • Which contracts deviate most from standards
   • Most common deviations across the portfolio
   • Priority queue for renegotiation based on risk × contract value

Playbook Structure

A playbook should define positions for key provisions. Example structure:

Provision: Limitation of Liability
  Preferred: Mutual cap at 12 months of fees; carve-outs for IP infringement and confidentiality breach
  Fallback: Mutual cap at 24 months of fees
  Non-negotiable: Must have mutual cap; unlimited liability is rejected

Provision: Indemnification
  Preferred: Mutual indemnification for third-party IP claims and confidentiality breach
  Fallback: One-sided indemnification acceptable if capped
  Non-negotiable: Must include IP infringement indemnity

Provision: Termination for Convenience
  Preferred: Either party, 30 days written notice
  Fallback: 60-90 days notice acceptable
  Non-negotiable: Must have termination for convenience right

Quality Checks

After generating any output:

• [ ] Every contract in the corpus appears in the inventory — none skipped
• [ ] Extracted dates are validated (no future effective dates on executed contracts, expiration after effective)
• [ ] Parent-child relationships are logically consistent
• [ ] Risk scores are justified with specific clause references
• [ ] Renewal deadlines account for notice periods (deadline = expiration minus notice period)
• [ ] Financial figures are in consistent currency and units
• [ ] Playbook deviations cite the specific section and language from both the contract and the playbook

Troubleshooting