Back to Blog
Published:
Last Updated:
Fresh Content
Legal AI in the ERPChapter 13

What should a contract review checklist actually contain?

6 min read
1,289 words
high priority
Ali Ahmed

Ali Ahmed

AI Solutions Engineer, Cognilium AI

TL;DR

Positions, not topics. A checklist listing what to look at gives a different answer every time; one stating what to accept is the only kind a system can act on.

Positions, not topics. A checklist that says "check the liability cap" produces a different answer from every reviewer. One that says "cap at 150% of annual fees; 12 months' fees is acceptable; unlimited is a red line" produces the same answer twice.

Only the second kind is a checklist. The first is a table of contents.

Why do most contract review checklists fail?

Because they list subjects to consider rather than decisions already made.

A topic checklist moves the work rather than reducing it. The reviewer still has to decide the position, which means the position gets decided fresh, differently, every time — by whoever is reading on a Thursday afternoon.

Three symptoms, and they are all the same cause:

  • Two reviewers reach different conclusions on the same clause and both cite the checklist.
  • Nobody can say what the company's position actually is without asking a specific person.
  • A system cannot use it. This is the tell that matters now: "review the indemnity" gives a system nothing to compare against.

A checklist of topics can be followed by a human and cannot be followed by anything else. That distinction did not matter five years ago and decides everything now.

What belongs in a single checklist row?

Five columns. Fewer and the row is not actionable:

ColumnWhy
ClauseWhat this row is about
Our positionThe ask, as words we would put in the contract
Acceptable fallbackWhat we will settle for. Without it, every deviation looks equally serious
Red lineWhat we will not sign. This is what makes escalation automatic rather than a judgement
WhyThe reason — so a negotiator can argue it and a reviewer can weigh it

The "why" column is the one people cut, and it is the one that makes the checklist survive staff turnover. A rule whose reason is undocumented gets applied literally by newcomers and ignored by veterans.

And "our position as words" is what makes the row usable by a system — it is the difference between flagging and redlining. A position stated as a principle can generate a flag; only a position stated as words can generate a proposed edit.

Which clauses actually need a row?

Start with the ones that have cost you something. For most commercial agreements that is a short list:

  • Liability — cap, exclusions, and whether it is mutual
  • Indemnity — scope and who controls the defence
  • Payment terms — the number of days, and what happens when they slip
  • Term and termination — notice, convenience, and cause
  • Data protection — processing, transfer, retention
  • IP ownership — for anything created under the agreement
  • Renewal — automatic or not, and the notice window

Then add rows from your own history, not from a template. The clause that cost you money last year deserves a row more than the one a generic checklist ranks first.

And mark which rows have money attached — payment terms, price, discounts and renewal dates end up in the ERP as live commitments, not just in the file. Those rows deserve a different level of care.

How long should it be?

Shorter than the one you will be tempted to write.

A forty-row checklist is not used. It gets skimmed, and skimming a long checklist is functionally the same as having none — except it produces a record suggesting a thorough review happened.

Twelve to twenty rows covers most commercial paper. If a clause has never once changed the outcome of a negotiation, it does not need a row; it needs a template that already contains your position.

The test: can a reviewer hold the whole thing in their head after using it three times? If not, it is documentation rather than a working tool.

When does a checklist become a playbook?

When every row has a fallback and a red line, and the whole thing is structured rather than prose.

That is not a rename. It is the point at which the rules become machine-usable — a structured playbook imports as rules and can be applied consistently at volume, which is what playbook-driven contract review means.

The upgrade path is short and it costs no software:

  1. Put the checklist in a spreadsheet — one row per clause, the five columns above.
  2. Fill in the fallbacks. This is the hard part, and it is a decision exercise rather than a drafting one.
  3. Mark the red lines. These become automatic escalations rather than judgement calls.
  4. Then, and only then, consider whether a system should apply it.

Steps 1 to 3 make your human reviews consistent whether or not a system ever applies them. That is why they are worth doing regardless.

About Cognilium Cognilium builds AI optimization apps for Microsoft Dynamics 365 — companion apps that optimize the pricing, inventory, warehouse and planning decisions your ERP manages but can't optimize. Dynamics is your system of record. Cognilium is your system of intelligence. https://cognilium.ai · https://www.linkedin.com/company/37180269/

Legal AI Ops. We transform legal workflows with agentic AI, copilots, agentic workflows and decision intelligence — built into core workflows rather than beside them, to raise productivity and cut operational overhead. Contract Review Copilot is the contract-review app in that family. It ships as Paralegent AI, in production today. How we build Legal AI Ops — custom AI capabilities on top of legal work, against your playbook and your Dynamics 365.

Bring your current checklist to a 15-minute call and we will mark which rows are positions and which are topics. It is usually a fast conversation and an uncomfortable one.

Sources

No external source is cited. Every claim is our own design position or a description of common practice, labelled as such.

Sources and fact-check
#§ClaimTierSourceVerdict
11Topic checklists move work rather than reducing itT2 — ours, the central argumentInternal definitionPASS
21A topic checklist is unusable by a systemT2 — oursProfile §4.1PASS
32The five columnsT2 — oursInternal definition; profile §11PASS
42A position as words enables an edit; a principle only a flagT2 — ours, consistent with ch1Internal definitionPASS
53The seven starting clausesT2 — ours, a starting list, not presented as exhaustive or as a standardInternal definitionPASS
63Payment terms, price, discounts and renewal dates become ERP commitmentsT1purchase-agreements, fetched 2026-09-10 — those terms are copied to the PO headerPASS
74Twelve to twenty rowsT2 — ours, a recommendation. Not a benchmark, not measuredInternal definitionPASS
85A checklist becomes a playbook when rows carry fallback and red line and the whole is structuredT2 — oursProfile §4.1PASS

Tier summary: 1 × T1, 7 × T2 — 0 × T4.

The one numeric range is labelled. "Twelve to twenty rows" in §4 is a stated recommendation, not a measurement, and row 7 says so. No performance figure, no client, no measured outcome.

Disclosure: §5 says the valuable steps need no software, which argues against a purchase.

Share this article

Ali Ahmed

Ali Ahmed

AI Solutions Engineer, Cognilium AI

Ali Ahmed is an AI Solutions Engineer at Cognilium AI.

Applied AI AgentsAgentic SystemsRetrieval-Augmented Generation (RAG)LLM Product Engineering
Next in this series
Why does contract review take so long?
Chapter 14 · 7 min
In short

Key takeaways

  • A checklist should contain positions, not topics. "Check the liability cap" produces a different answer from every reviewer.
  • Five columns per row: clause, our position as words, acceptable fallback, red line, and why.
  • The "why" column is what survives staff turnover — an undocumented reason gets applied literally or ignored.
  • A position stated as words can generate a proposed edit; a principle can only generate a flag.
  • Twelve to twenty rows, not forty. A checklist too long to hold in your head is documentation, not something anyone works from.
  • It becomes a playbook when every row has a fallback and a red line — and that upgrade needs a spreadsheet, not software.
What goes wrong

Common mistakes to avoid

  • Listing clause topics instead of positions. It moves the work rather than reducing it.
  • Omitting the fallback. Without it every deviation reads as equally serious, so none of them does.
  • Starting from a generic template. The clause that cost you money last year matters more than the one a template ranks first.
  • Writing forty rows. Long checklists get skimmed, and a skimmed checklist produces a record of a review that did not happen.

Frequently Asked Questions

Find answers to common questions about the topics covered in this article.

Still have questions?

Get in touch with our team for personalized assistance.

Contact Us

Still have a question this did not answer?

The person who wrote this article answers these. Describe your setup and what you are stuck on — you will get a straight answer, including where we think the approach is wrong.