An application of Sourceline. It reads every lease in the portfolio and answers what each one obliges: option windows, rent reviews, make-good and outgoings, reconciled to the document rather than the summary. Every answer shows the clause it came from.
For owners, directors and CFOs of private property groups. Private Australian hosting is available. Where it is selected, your data stays in Australia.
Managing agents administer the leases day to day. The owner sees a rent roll and a set of invoices. Between the lease and the summary sits a set of dated obligations. Unless someone reads the lease itself, each one is taken on trust.
An option to renew has a window. It opens and closes on dates set by the lease, often months before expiry. A window missed is a tenancy lost or a rent reset on the landlord's terms instead of yours.
A market review has to be initiated by a date. If the trigger passes without a notice, the review is forgone for the period. The rent roll will not tell you a review was available and not taken.
Make-good obligations fall due at expiry. Outgoings recoveries are defined by the lease, not by whatever the invoice says. CPI increases apply on an anniversary. Each is a clause with a date attached.
The system reads the lease, then reads the record of what was sent and when, and shows where the two do not line up. It does not accuse anyone. It shows you what is currently unverified.
Every obligation the system surfaces carries a reference to the lease and the page it came from. Your people verify against the source, not against a confidence score.
Every option window, rent review trigger, CPI date, make-good and outgoings obligation across the portfolio, each with the lease, the clause and the dates it runs between.
Page-referenced to the source leaseThe dated obligations ordered by what falls next, so the windows about to open or close are visible before they pass rather than after.
Updated as new leases are onboardedWhere the record of what the agent sent does not match what the lease requires, flagged for your review. The system drafts the question; your people decide.
Uncertain reads are marked uncertainMany AI pipelines flatten documents into plain text at ingestion. Tables become paragraphs. The clause number detaches from the clause. The chain back to the source page disappears, replaced by a confidence score.
That is acceptable for a chatbot. It is not acceptable for an option deadline, a payment claim, or anything a lawyer or adjudicator might one day read.
Where documents carry the decision, the system runs on Sourcetrace, our document-structure engine. It preserves layout, tables, annotations and page evidence before anything reaches a model.
Every output cites its source document and page. No orphan facts.
Your environment or private Australian hosting where selected. No training on your data.
Named human review points before any output is relied on.
A log of what the system read, produced and flagged.
Uncertain answers are marked uncertain. Confidence is reported, not performed.
AUD 5,000, all in.
Your board sends its questions ahead. Every one comes back answered in writing. Then ninety minutes in the room, and you leave with decisions, not a presentation.
One week, end to end.
Book the Board SessionAUD 9,500, fixed.
Pick the one job you think AI could do. A contract read. A report produced. A cycle shortened. We run it on your own material for two weeks. It works, or you get a written account of why it does not.
Start a ProofFrom AUD 35,000, quoted after a Proof.
We build the system, fit the controls, and run it. One proposal, one number, and the running of it is in the number.
Talk about a Build