Lease Break Notice Validity Checker

Leasing counsel often receive a break notice late in a tight window. The notice may look complete, yet fail on a technical point that the lease or the governing law treats as fatal: wrong address for service, missing required wording, premature or late timing, or a method of service the lease does not allow. Once the break date passes, curing a defective notice is frequently impossible.

This use case describes an LLM-assisted checker that compares a proposed or received break notice against the operative lease and a jurisdiction ruleset. The model surfaces candidate defects with citations to the clause or rule that appears to apply. Counsel remains the decision-maker on whether the notice is valid, whether a defect is remediable, and what remedial step to take.

The checker is built for quality control at the vacate stage of a property lifecycle. It does not replace legal advice, does not file or serve documents, and does not invent lease terms or local procedure when those inputs are absent.

Three inputs are mandatory. If any one is missing, the system returns empty output rather than a speculative review.

Break notice. The draft or served notice text, plus any cover correspondence that forms part of the service package. Prefer the full PDF or Word file so headers, addressees, and signature blocks are visible. A truncated excerpt is not enough when the lease requires specific form language.

Lease extract. At minimum: break clause, service of notices clause, definitions of landlord and tenant, and any schedule that sets addresses or process agents. If the lease has been varied, include the variation that affects break rights or service. Without the operative wording, the model cannot test compliance.

Jurisdiction rules. Counsel-supplied notes or a curated rules pack for the relevant country, state, or court practice covering statutory form requirements, deemed-service timings, and any case-law gloss the firm treats as binding for this asset class. The model must not fill gaps from general training knowledge when the rules pack is empty.

Optional enrichments improve precision without becoming substitutes for the three mandatory inputs: a chronology of prior notices, proof-of-service evidence (postal receipt, courier log, email headers), and a short matter memo stating who is serving whom and under which break option.

Counsel (or a paralegal under supervision) uploads the notice and lease extract, selects the jurisdiction rules pack, and states the intended break date and party roles. The model then walks a fixed checklist rather than free-form commentary.

1. Party and capacity. Confirm the notice names the correct landlord or tenant entity, including trading names and guarantors only where the lease requires them.
2. Break conditions. Map any preconditions in the break clause (vacant possession, rent paid up to date, no outstanding breaches, prescribed form) and mark each as stated, not stated, or not evidenced in the package.
3. Timing. Compare notice period, earliest and latest service dates, and any “time of the essence” language against the stated break date and the deemed-service rules in the jurisdiction pack.
4. Method and place of service. Check permitted methods (personal delivery, recorded post, email, process agent) and the contractual address for service against what the notice package claims was used.
5. Form and content. Flag missing mandatory phrases, incorrect references to clause numbers, or optional break options selected in the wrong combination.
6. Defect severity hint. Label each finding as potentially fatal, potentially remediable, or informational, with the explicit caveat that severity is a prompt for counsel review, not a determination.

Output is a structured findings list: issue, supporting excerpt from the notice, supporting lease or rules citation, and an open question for counsel where evidence is incomplete. The interface should make it easy to export findings into the matter file and to re-run the check after a revised draft.

The model’s job is defect spotting under constraint. Typical flags include:

• Service addressed to a historic registered office when the lease nominates a different notice address.
• A notice period counted from the wrong trigger (date of posting versus deemed receipt).
• Conditional language that undercuts an unconditional break right.
• Failure to exercise the correct break option among several in the same lease.
• Attachment of the wrong plan or demise description where the lease makes identification a condition of valid exercise.

Counsel still decides validity. Local courts and arbitral practice can treat some technical defects as saved by statute or by the recipient’s actual knowledge; others remain strict. Only counsel can weigh those doctrines against the facts, client risk appetite, and opposing counsel’s likely challenge. The checker must never emit a binary “valid” or “invalid” stamp that could be mistaken for a legal opinion.

Human-in-the-loop controls should be explicit in the product behavior: findings stay in draft until a named reviewer accepts, edits, or discards them; overrides are logged; and empty-input runs produce no findings rather than generic advice.

Missing or wrong lease version. A checker pointed at an outdated lease will miss variations that shortened notice periods or changed service addresses. Treat version control as a precondition, not an afterthought.

Jurisdiction pack drift. Statutory amendments and leading cases change deemed-service and form rules. Firms should version the rules pack and record which version applied to each review.

Ambiguous PDF text. Scanned notices with poor OCR quality can omit critical words. Prefer native digital files; when only a scan exists, require human confirmation of OCR before relying on findings.

Cross-border portfolios. A single prompt that mixes English-law and other jurisdictions without separate rules packs will produce unreliable timing analysis. Run one jurisdiction per review.

Over-trust. Teams under deadline pressure may treat a clean findings list as clearance. Training and UI copy should state that a nil-defect report means “no defects detected against the supplied inputs,” not “the notice is safe to serve or rely on.”

When the notice, lease extract, or jurisdiction rules are not provided, return empty output. Do not invent clause numbers, statutory citations, or customary practice to complete the checklist.

Break-notice quality sits beside other vacate-stage controls. Condition and dilapidations evidence often travels with the same matter timeline; see Dilapidations Assessment via Computer Vision. Once a break is progressing, document packs for handover and deposit accounting benefit from structured generation; see Move-Out Document Automation. Where the asset will be re-let, screening and match quality for incoming interest is a separate quality outcome; see Prospective Tenant Matching Model.

Used together, these workflows keep technical notice risk, physical condition evidence, and re-letting readiness on parallel tracks without collapsing counsel judgment into a single automated score.