Maya buys a house.
Her lawyer gets the moving parts.
One agreement turns into people to contact, conditions to track, dates to meet and documents to get right. ConveyIT helps the firm run that work.
14 Kōwhai Lane, Exampletown
“I just want the keys on time.”
Maya Reed · example purchaser
Reed — property purchase
A matter is the firm’s file for one piece of client work.
Who is buying? Which property? Which lender and other lawyer?
What must happen, who owns it, and when is it due?
Use the right legal content, facts and firm’s letterhead.
Coordinate checks, signing, settlement and official title work.
ConveyIT sells an organised way through a property transaction. The screens are the delivery mechanism.
The know-how is the product.
ConveyIT sits alongside a law firm’s broader systems. Its strongest asset is the property-law work encoded in its workflows and document rules.
A reusable recipe for the work, adapted to a real matter.
Approved legal content assembled using facts and choices.
Branding, settings and overrides without losing shared expertise.
Gibson Sheat origins, 1999 → Standalone business, 2005 → Cloud product today
The strategic question: why should a firm pay for ConveyIT as its practice system adds more automation?
The market is already adding AI. What could make ConveyIT different?
Matter AI and drafting marketed in NZ.
AI launched August 2026; first capabilities: time capture and business intelligence.
October terms describe AI as coming soon.
Our hypothesis: dependable NZ property workflows, maintained legal content and less retyping between systems. Test it with buyers. A generic assistant alone is a weak reason to buy.
How does ConveyIT make money?
Published terms describe onboarding/customisation and minimum/transaction-based charges. The document also contains old technical requirements, so confirm current contracts and pricing. Ask Dean whether growth means more firms, more matters, premium features, or lower support cost.
The website is only the front door.
It is already a cloud product. Behind it are legal rules, Windows authoring tools, document engines and connections that practices rely on.
OneLaw support in August 2026 described matter-resave triggers, five-minute processing and field-mapping issues. A better product should make ownership and failed transfers visible.
A successful rebuild preserves behaviour people depend on. Correct dates and documents matter more than matching the old screen.
What Hashim and Werner already have to work from
Your Microsoft SSO work was reported live on 21 January 2025. Greg’s latest branch contains real Angular/.NET implementation, including identity, contact writes and read-only matters/API. His knowledge maps are useful; the technical foundation needs your review. Full workflows, sync and cutover remain incomplete.
The checked-out code is not a complete picture of production. Ask for current schema, deployment, volumes and anonymised document examples. Dashboard code suggests a performance investigation; it does not prove the live bottleneck.
A contract arrives.
Which date enters the matter?
Useful AI turns a document into facts someone can check. Try the review, then see where the accepted fact can go.
Source / page 1
Settlement date
20 November 2026
The agreement also contains special conditions.
Source / page 4 · clause 8
The parties agree to change settlement to 23 November 2026.
Two source values disagree. A person must resolve the legal position.
Proposed matter fact
20 November
Candidate · page 1Nothing has been written to the matter.
Extract → show the source → review → reuse the accepted fact. Deadline calculations stay in tested rules.
Richard’s June 2025 outline already proposed this reviewed intake flow, plus LIM summaries and title-search research. Your experiments stayed in a sandbox.
Five requests.
Five different questions.
Pick a feature. Each needs a user outcome, a dependency and a way to prove it works.
Identity → matter → workflow → correct document. Test the difficult rules early.
Reviewed intake or a focused client form. Measure time saved after review.
Grow signing, forms and integrations in pilot-informed order.
Ask for examples before estimating features. “LINZ integration” could mean a lookup—or a much larger legal transaction flow.
We already use LINZ
in AlarmWatch.
The site map combines LINZ photos and geographic data to locate a site, draw its surroundings and give the AI draft useful context.
What you can say: “We've used LINZ's mapping services for aerial photos, building outlines, addresses, parcels and roads in AlarmWatch. We know how to fetch and combine that data. For ConveyIT, we need to pin down which property or Landonline task they want to connect.”
The important boundary: parcel data can include title references. That is different from obtaining an official title document, checking ownership or registered interests such as mortgages, or submitting a Landonline dealing.
If they ask how it works
Two services: LINZ Basemaps supplies image tiles, which our server joins into a picture. The LINZ Data Service supplies layers of shapes and facts as JSON, a structured data format. They use separate API keys.
Our code caches results, keeps source credits, accounts for image scale and rotation, and can fall back to other map providers. Photos and outlines can be out of date, so we check imagery dates and alignment rather than assume the map is current.
Useful next question: do they want property identification and context, an official title/search document, or a step inside a Landonline transaction?
One product.
Clear places for the rules.
Start with a modular application a small team can own. The web interface, business rules and external connections have separate jobs.
Modernise the rules in measured steps. Keep SQL Server initially; replacing every procedure does not need to block the first useful release.
The practical tool answer, if David asks
Candidate: Angular/TypeScript, ASP.NET Core/.NET 10, SQL Server, EF Core for appropriate new data access, Git and automated checks. Use reviewed examples to test dates, tenant access and generated Word/PDF output. Open XML is worth investigating for DOCX generation; choose rendering and editing after fidelity proofs.
Keep Microsoft SSO. Run reliable background jobs for documents, messages and sync. Put each provider behind an adapter. Choose hosting, storage and an AI provider after inspecting operational needs and data-handling terms.
Move one firm.
Know who owns the writes.
Both products can run during rollout. Each firm’s records and background jobs still need an agreed owner. Rehearse a pilot below.
Existing product
Authoritative writer
New product
Authoritative writer
Migrating a practice includes more than its database. Include files, legal definitions, permissions, external IDs and job ownership.
Clarify “move clients individually”: law-firm customers, end clients, or matters? Decide whether active matters move or finish in the old product.
They need a business case.
We need the right starting point.
Use the meeting to connect their business goals to a pilot, its proofs and the people who can approve it.
Before you go: Greg should bring the exact proposal they received. David Deacon is tomorrow’s consultant; David Sarginson is the Gibson Sheat contact in the older AI emails. Richard is a valuable follow-up source for workflows and integrations.
A useful 2-hour shape
15 min: goals and board decision. 25: two real matters. 25: architecture, tools and developers. 25: feature outcomes. 20: migration and pilot. 10: owners and next actions. Use an extra hour for staff and integration walkthroughs.