A spellchecker for rules: checking documents as you write

The cost of checking too late

Most organisations discover non-compliance at the end of the process: during a final legal review, a tender evaluation or, worse, after publication. By then fixing it means rework, delays and sometimes lost opportunities. The rules were known all along. They simply were not present while the document was being written.

Like a spellchecker, but for rules

A spellchecker does not wait until a text is finished. It flags a problem the moment it appears and shows exactly where it is. This YasIQ application brings the same experience to rules. It understands what is being written or reviewed, checks it against an identified and governed Rulebook, and shows:

• where the problem is in the document,
• which rule applies,
• what the rule requires, and
• why it matters.

Reading the document as a whole

Compliance is rarely about a single sentence. The check considers the document as a whole: whether required content is present, whether it appears in the right position, and whether the document is coherent, so that a commitment in one section is not contradicted in another. Claims can be verified against the evidence that supports them.

AI intelligence, without AI authority

The rules themselves are not invented by the AI. The Rulebook is identified, versioned and owned by the organisation: its policies, its contractual playbook, a tender’s requirements or the applicable regulation. AI is used to understand language and context; it has no authority over what the rule says. That separation is what makes the result trustworthy.

Verification you can prove

Every check produces a verification record: which version of which rules was applied, what was found, how it was resolved and who approved the final document. When a regulator, a client or an auditor asks how compliance was assured, the answer is a record, not a recollection.

Where it applies

• Tenders and bids: confirming every mandatory requirement is addressed before submission.
• Privacy and data protection: checking that notices contain the information the law requires.
• Contracts: flagging clauses that depart from the organisation’s positions or cross its red lines.
• Corporate communication and policies: keeping claims and internal documents aligned with the rules that govern them.

Local by design

The documents being checked are often confidential. The application can run with local AI, so sensitive content does not need to leave the organisation’s environment.

Watch on YouTube · 2-minute explainerThe Spellchecker for Rules: Check Documents as You Write | Governed AIYasIQ channel · opens in a new tab

Key takeaways

• Rules are usually checked at the end, when fixing problems is most expensive.
• Checking against a governed Rulebook while writing shows where the problem is, which rule applies and why.
• AI understands the text; it never has authority over the rules.
• Each check leaves a verification record that can be shown to auditors and clients.

Frequently asked questions

What is a Rulebook?

An identified, versioned set of rules owned by the organisation, such as internal policies, a contract playbook, tender requirements or applicable regulations, against which documents are checked.

Does the AI decide what is compliant?

No. The AI interprets the document and points to the relevant rule; the rules are defined by the organisation, and people review findings and approve the final document.

Can confidential documents be checked without sending them to external AI?

Yes. The application can run with local AI within the organisation’s environment.