Skip to content
Hidden In NumbersHidden In NumbersAI-Powered Document Intelligence
All posts
anomaly detection

Date Inconsistencies: The Silent Contract Killer

Ibrahim ArbiJune 7, 2026 8 min read

Dates seem like the simplest thing to get right. They are just numbers, after all. But date inconsistencies are among the most common and consequential errors in professional documents — and among the hardest to catch manually.

The types of date inconsistency

Date problems fall into a few recurring categories.

**Impossible sequences.** An effective date that falls after an expiration date. A report period that ends before it begins. A signature date that precedes the document's stated drafting date.

**Conflicting references.** The cover page says Q3 2024; the body says Q2 2024. The contract header says 15 March; the payment schedule says 14 March for the same event.

**Format inconsistencies.** Mixing DD/MM/YYYY and MM/DD/YYYY in the same document. This ambiguity makes dates like 04/05/2024 genuinely unresolvable — is it 4 May or 5 April?

**Stale boilerplate.** A template with dates from a previous version that were not updated. This is extremely common in policy documents and contracts that are adapted from prior work.

Why dates cause legal problems

Courts and regulators treat dates literally. An effective date of 1 January in a contract that was not executed until 15 January may be unenforceable, or it may create obligations that were never intended. A reporting period stated incorrectly can misrepresent financial results. A policy dated in a prior year may not apply to a current incident.

Insurance claims, employment disputes, and procurement challenges have all turned on nothing more than a date that was copied incorrectly from a template.

Manual detection is unreliable

A document with 20 date references requires 190 pairwise comparisons to check every relationship. Humans do not do this. They check the cover, glance at the signature block, and move on.

The problem is not carelessness — it is cognitive load. Full coverage requires a systematic method.

Automated approaches

Automated date detection works in two passes. First, all date expressions are extracted and normalised to a standard format. This handles the variety of ways dates appear: "the first day of January 2024," "01/01/24," "January 1st," "2024-01-01."

Second, the normalised dates are checked against logical rules: effective dates must precede expiration dates, document periods must be internally consistent, and the same event should not have two different stated dates.

The output is a list of conflicts with the exact passages that generated them, so a reviewer can evaluate each one in context.

Format standardisation as a preventive measure

The simplest mitigation is a house style for dates. Picking one unambiguous format — ISO 8601 (YYYY-MM-DD) or a spelled-out month (15 January 2024) — and enforcing it eliminates the ambiguity class entirely.

But even with a house style, the sequencing and consistency checks remain necessary. A template carries the wrong year regardless of what format it uses.

Before you sign or publish

Run a date check as the final step before any document is executed or published. It takes seconds and protects against a class of error that is easy to introduce, invisible to casual review, and expensive to fix after the fact.

Try it on your own document

Scan a Document
Hidden Signal Engine™ Online

Discover What Your Documents Are Hiding

AI-powered document intelligence that uncovers contradictions, anomalies, missing references, and hidden patterns — in seconds.