AI for Contract Intelligence Capstone Exercise — AI for Contract Intelligence

What you'll learn

Apply skills from across this course — extraction, obligation tracking, risk clause identification, negotiation prep, and governance — in a single realistic contract review scenario
Produce a concrete, role-relevant deliverable using an AI-assisted contract review workflow
Self-assess your output against the verification and escalation standards this course has established

This course has covered AI across the contract lifecycle: CLM automation and workflow routing, data extraction, obligation tracking, risk clause identification, negotiation prep, and governance and accuracy controls. The capstone brings those skills together in a scenario a contract manager encounters routinely: a new vendor agreement lands on your desk, and you need to process it through an AI-assisted workflow while catching exactly the kind of errors this course has documented — a misdated obligation, a missed non-standard clause, an extraction pulled from the wrong section of the document.

The exercise is designed to test not just whether you can produce the outputs, but whether you can identify the verification gaps an AI tool will not flag itself. In contract work, an unverified AI output that drives a missed renewal deadline or an unflagged liability exposure is a real financial and operational risk. The goal is to move fast without moving recklessly — and to remember throughout that substantive contract risk judgment belongs to qualified counsel, with AI and this course's frameworks providing decision support, not a replacement for that judgment.

Capstone Exercise

New Vendor Agreement: Extraction, Obligation Register, and Risk Escalation

Context

You are a contract manager at a mid-size logistics company. A new three-year software licensing agreement with a warehouse management system vendor has just been countersigned. The agreement is 34 pages, includes two exhibits (a pricing schedule and a service level exhibit), and contains a limitation of liability clause that caps damages at 12 months' fees except for a carve-out defined by reference to a 'Excluded Claims' definition located in Section 1 (Definitions). The agreement auto-renews annually unless either party gives 90 days written notice, and includes a most-favored-nation pricing clause tied to the vendor's other customers of similar size.

Your Task

Working from the scenario description above, produce four deliverables. (1) A structured extraction record listing the key fields you would expect an AI extraction tool to pull — parties, term, renewal notice window, governing law — and flag any field where you would want to verify against the source document before trusting it, and why. (2) An obligation register entry for the most-favored-nation pricing clause, including what would trigger it, what a contract manager would need to actively do to enforce it, and why this is a conditional obligation type this course has flagged as high-risk for AI extraction. (3) A risk escalation note for the limitation of liability clause, explaining specifically why the cross-referenced 'Excluded Claims' definition makes this a clause that should not be trusted to AI classification alone, consistent with the risk clause identification lesson. (4) A one-paragraph reflection (100 to 150 words) on which of this course's verification disciplines — QA sampling, conditional-obligation review, false-negative risk clause sampling, or governance escalation rules — would have been most important to apply to this specific contract, and why.

Your notes (optional)

Deliverable

A structured extraction record with verification flags, an obligation register entry for the MFN clause, a risk escalation note for the limitation of liability clause, and a 100-to-150-word reflection identifying the most important verification discipline for this specific contract.

Quick check

The obligation register entry for the most-favoured-nation pricing clause has to record more than the fact that the clause exists. What does this capstone say is missing if it does not?

Select one answer.

Key takeaways

AI contract intelligence tools are a genuine accelerant across the full contract lifecycle — extraction, obligation tracking, risk classification, negotiation prep — but every output that drives a financial, compliance, or renewal decision must pass through the verification standard this course has built lesson by lesson.
Conditional, cross-referenced clause language — an obligation triggered by a defined term located elsewhere in the document, a risk clause carve-out defined in a separate section — is the single most consistent failure pattern across extraction, obligation tracking, and risk classification covered in this course.
A risk-tiered QA sampling protocol, applied consistently rather than left to individual judgment, is what turns catching an error like a misclassified liability clause into a designed outcome rather than a fortunate accident.
The limit of AI in contract work is not capability but verification: an AI tool can extract, classify, and draft in seconds, but the contract manager or qualified counsel who relies on that output remains responsible for confirming it before it drives a real decision.