Back to Blog
Published:
Last Updated:
Fresh Content
Legal AI in the ERPChapter 6

Pre-signature or post-signature — which contract AI do you actually need?

7 min read
1,529 words
high priority
Ali Ahmed

Ali Ahmed

AI Solutions Engineer, Cognilium AI

TL;DR

Two different products sold under one name. Pre-signature reads a draft against your rules; post-signature watches what you agreed. Most teams buy the first.

They are two different products sold under one name. Pre-signature reads a draft against your rules before you sign. Post-signature watches what you already agreed and tells you when reality drifts from it.

Most teams buy the first and discover they needed the second.

What is the split, exactly?

It is a split by when the money is at risk, and the two halves share almost no machinery.

Pre-signaturePost-signature
The question"Should we agree to this?""Are we doing what we agreed?"
The inputa draft, and your playbooka signed set of obligations, and live transactions
Runsonce per contractcontinuously, for the life of the agreement
Fails byapproving a bad clausenever noticing a breach
Sold ascontract review · redlining · playbook reviewobligation management · contract intelligence · CLM

The vocabulary is the trap. Vendors on both sides call it contract intelligence, so a buyer comparing two products may be comparing a reader and a monitor as though they were rivals.

Ask which question the product answers. That single question sorts the market in a minute.

What does pre-signature actually do?

Reads a draft against your rules and reports the gaps. That is playbook-driven contract review — the clause is measured against your standard position, your fallback and your red line, and a reviewer decides.

Its value is concentrated in one moment and it is genuinely high: the cheapest time to fix a term is before anyone signs it. Its limitation is the same fact. Once signed, it has nothing left to say. The document is filed and the system's job is finished.

This is what almost every "AI contract review" product is, and it is what we ship — Paralegent AI is a pre-signature reader, in production today, and one of Cognilium's AI optimization apps for Microsoft Dynamics 365. We build these on request against a customer's own playbook.

And it is not the half most procurement teams are actually losing money on.

What does post-signature do that pre-signature cannot?

Watch the record. And in procurement, the record is the ERP — which is where the interesting failures live.

Microsoft's purchase agreement is a live commitment, not a filed document. It carries the operative terms, and Microsoft is explicit that they reach the transaction:

"Information from the terms for the agreement, such as the payment terms, delivery terms, and delivery address, is then copied to the header of the PO… the prices and discounts from the purchase agreement are used for those lines."

So the agreement is doing work every time somebody buys something. And it can stop doing that work silently:

"If the price is changed on the order line, the link to the commitment is broken. If the link is broken, the order line doesn't contribute to the fulfillment of the commitment."

No error. No alert. The volume you promised in exchange for the discount simply stops accumulating.

That is a post-signature failure and no pre-signature review can see it. The contract was read correctly. Months later, an ordinary price edit detached it.

So which do you need?

Answer one question: what has actually cost us money in the last two years?

  • A term we should never have agreed topre-signature. The review is the fix.
  • A term we agreed and then did not getpost-signature. Reading the draft again changes nothing.
  • We do not knowpost-signature, and start by reading your own agreements. Not knowing is itself the post-signature answer.

The honest pattern we see: teams buy pre-signature because it demonstrates well — you can watch it find a bad clause in a meeting — and their real leakage is a renewal nobody tracked, a rebate nobody claimed, or a discount tier nobody accumulated toward.

Neither half is the wrong purchase. Buying the wrong one first is the expensive part.

Can one system do both?

Parts of it, and the parts that do not transfer are the parts that matter.

The reading machinery is genuinely shared. Clause identification, playbook matching and risk scoring work the same whether the document is a draft or an executed agreement. A pre-signature reader pointed at a signed contract will extract obligations perfectly well.

What does not transfer is everything after extraction:

  • Post-signature needs to run forever, on a schedule, against changing transactions. Pre-signature runs once and stops. That is a different operational shape, not a bigger version of the same one.
  • Post-signature needs write access to nothing and read access to everything — orders, invoices, receipts, price changes. Pre-signature needs a document and a playbook.
  • Post-signature alerts a person about something already true. Pre-signature advises a person about something not yet agreed. One is monitoring; the other is review.

So the honest answer: extraction is shared, monitoring is a second build. A vendor claiming one product covers both is usually selling the reader and calling the extracted obligations a monitor.

Why does the ERP decide which half matters?

Because it holds the obligations that have money attached, and it enforces them on rules a legal team never sees.

Microsoft's own agreement policies are the clearest illustration. Max is enforced caps the total against the commitment. The price and discount is fixed policy breaks the link on an edit. And an agreement can be attached only at order creation — "You can't select a purchase agreement after the PO has been created."

Those are contract terms living as configuration. They govern quietly, they are changed by people who have never read the contract, and they are invisible to any product that treats a contract as a document.

Post-signature contract intelligence in an ERP means watching the record, not re-reading the file — and it is the half almost nobody publishes about.

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.

Not sure which half you need? Bring two years of contract problems to a 15-minute call and we will sort them into pre- and post-signature with you. The answer is often not what people expect.

Sources

Sources and fact-check
#§ClaimTierSourceVerdict
11The pre/post split, and that both are marketed as contract intelligenceT2 — ours, a market reading_research/2026-09-09/01-RESEARCH-A… §2, nine SERPs readPASS
22Pre-signature reads a draft against playbook rules and stops at signatureT2 — ours + capabilityParalegant_TECHNICAL_PROFILE.md §1PASS
32Paralegent AI is a pre-signature reader, in productionT2 — capability. Founder-locked canon for status/dynamics-365/optimizers; 04-positioning-now.mdPASS
43The copied-terms passage, quoted wholeT1 — verbatimpurchase-agreements (ms.date 2026-09-08), fetched 2026-09-10PASS — load-bearing
53The broken-commitment-link passage, quoted wholeT1 — verbatimSame page, Policies for purchase agreementsPASS — load-bearing
63No error is raisedT2 — ours, bounded: Microsoft documents the link breaking and documents no error. Stated as behaviourSame policyPASS
74Buy against loss history; the pattern we seeT2 — ours, deliberately unquantified. No client, no countInternal definitionPASS
7b4aExtraction machinery transfers between the halves; monitoring does notT2 — ours, an architectural argumentInternal definition; profile §1PASS
85Max is enforced; price-and-discount-fixed; attachment only at PO creationT1 — verbatim ×3Same pagePASS — load-bearing
95Contract terms live as configuration, changed by people who never read the contractT2 — ours, an argumentInternal definitionPASS

Tier summary: 5 × T1 (all verbatim), 5 × T2 — 0 × T4.

Absence claims: one, row 6, bounded to the fetched page and phrased as observed behaviour rather than as a claim about Microsoft's complete documentation.

Disclosure: no client, no count, no measured outcome. "In production today" is founder-locked canon.

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
When should contract AI escalate to a human reviewer?
Chapter 7 · 7 min
In short

Key takeaways

  • Pre-signature and post-signature are two products sold under one name. Pre-signature reads a draft against your rules; post-signature watches whether you are doing what you agreed.
  • The sorting question is which one the product answers: "should we agree to this?" or "are we doing what we agreed?"
  • Pre-signature has nothing to say once the contract is signed. Its value is concentrated in one moment, and that moment is genuinely the cheapest place to fix a term.
  • In procurement the post-signature record is the ERP, and it fails silently — a price edit can detach an order line from the commitment with no error raised.
  • Buy against your own loss history. A term you should not have agreed to is a pre-signature problem; a term you agreed and did not get is not.
What goes wrong

Common mistakes to avoid

  • Comparing a reader against a monitor as though they were rivals. Both are marketed as contract intelligence and they solve different problems.
  • Buying the half that demos well. Watching a system find a bad clause is compelling; the leakage is usually a renewal nobody tracked.
  • Treating a signed contract as filed. In an ERP it is a live commitment that prices real orders.
  • Assuming a broken commitment raises an error. It is documented, silent behaviour.

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.