Why "genuinely" is the operative word

Most conveyancers reading this have already tried an AI tool. Fewer can say what it saved them.

According to the Thomson Reuters Future of Professionals Report 2026, which surveyed 1,816 professionals across law, tax, accounting and compliance in 62 countries, 74% now use AI tools several times a week and 44% use them multiple times a day. According to the same report, 91% of respondents have felt some gap between what AI promised and what it delivered, and only 18% of organisations track return on investment at all.

The Australian picture is more cautious. According to the Victorian Legal Services Board and Commissioner's report on generative AI use, drawn from the 2025 Victorian Lawyer Census, 36.7% of respondents use AI tools in their legal practice, and over half of those use them daily or weekly.

"Currently, lawyers are mainly using AI to improve workflows and boost productivity."

Fiona McLeay, CEO and Commissioner of the Victorian Legal Services Board

triSearch.ai reads those two data sets together. Adoption is well advanced; the missing piece is that most firms cannot point to a task, a document type and a number of minutes.

Vague productivity claims have trained practitioners to be sceptical, and the scepticism is earned. Specificity is the fix, and it is what the rest of this article is for.

Where the hours go in a conveyancing file

Settlement is not the slow part. According to PEXA's third-quarter FY2026 update, the platform processed 935,000 transactions in the quarter, up 7.3% on the prior comparable period, with transfer volumes up 7.1% to 588,000 and national market penetration steady at 90%. The exchange itself is digital and fast.

The hours sit upstream. According to Thomson Reuters research, lawyers spend between 40 and 60% of their time drafting legal documents and reviewing contracts. In a conveyancing practice those documents are the contract of sale, the special conditions, the section 10.7 certificate, the strata report, the search bundle, and the stream of emails and letters that surround them.

The calendar is the other cost. A purchase matter runs for weeks because councils, lenders, counterparties and strata managers take time to respond. Software doesn't change their turnaround.

The two clocks measure different things. One counts the practitioner's effort in hours; the other counts elapsed days.

AI moves the first clock. It leaves the second alone, and any vendor who tells you otherwise is describing a different transaction.

A new line item arrived in July. From 1 July 2026, conveyancers and lawyers who provide designated services, including conveyancing to give effect to the sale or purchase of real estate, carry obligations under the Anti-Money Laundering and Counter-Terrorism Financing Act, according to Grant Thornton's summary of the Tranche 2 reforms. Those obligations include enrolling with AUSTRAC, running customer due diligence, keeping records for seven years and reporting suspicious matters.

Two months in, that work is real and it is landing on the same staff who were already stretched. In triSearch.ai's view, the intake step is the highest-value place to automate, because the data you capture at the front door now has to travel further than it used to.

Where AI genuinely saves time: five tasks

The pattern is consistent across all five. AI does the reading, extracting and first-draft work. You do the checking and the deciding.

Swipe sideways to see the full table →

1. Matter setup and client intake

Re-keying the same client name, address and property description into a form, a matter file and a search order is waste in its purest form. It also introduces the typos that later stall a settlement over a mismatched figure.

Digital intake forms that write directly into the matter remove that step. triSearch.ai's Intake Forms are built for leads and matters, capture the details from a new or existing client, and sync the data into triConvey.

Since July, the same intake moment is also where customer due diligence starts. Capture it once, in a structured field, and it feeds the AML/CTF record instead of being typed a second time for AUSTRAC purposes.

Chris Gibbs, triSearch's chief executive, said a matter creation demonstration condensed an hour's work into seconds. triSearch.ai would put it more carefully than that for a typical firm: the saving per matter is measured in tens of minutes, and the bigger win is the errors that never happen.

2. Contract and special-condition review

The first review of a contract of sale is a comprehension task. You are looking for the special conditions, the encumbrances on title, the planning controls in the s10.7, and anything that contradicts something else in the bundle. AI does that work well, provided it has been built on the right contract.

triSearch.ai's AI Contract Review was designed around the NSW Law Society 2022 Contract for the Sale and Purchase of Land. Upload the contract from the Quick Launch screen in One Flow and the side panel populates with flagged clauses and conditions.

The tool currently reviews the general contract information, the special conditions, the title search and the section 10.7 certificate, and it is available for NSW purchase matters. It is also a product in active development, according to triSearch's own site.

The tool highlights. It does not advise. The decision about whether a special condition is acceptable for your client, and what to say about it, is yours.

triSearch.ai's position is that a review tool earns its place when it reliably surfaces the items you would have found anyway, faster, so that your time goes to the judgement call rather than the hunt.

3. Reading long documents for a single answer

An 80-page strata report contains one paragraph about pets. A search bundle contains one line about a sewer easement. The manual method is to read the lot; the AI method is to ask.

Archie, triSearch.ai's assistant inside triConvey, summarises a matter in a few clicks and answers matter-specific questions from the documents on file. Gibbs said a specific by-law can be located in an 80-page strata report in seconds.

That matches what triSearch.ai hears from practitioners: the saving is largest on documents that are long, standardised and only partly relevant to the question at hand.

The limit is the document type. Scanned deeds, handwritten annotations and unusual instruments are where AI is most likely to be confidently wrong. Read those in full.

4. Drafting routine correspondence

Client updates, requisition letters, settlement notices and cover emails follow patterns. The pattern is the part AI is good at.

Archie drafts documents, clauses, memos, emails and letters, and you can adjust the tone, length or style before sending. The practitioner's job on each draft is the same as it would be for a junior's draft: check the facts, fix the phrasing, sign.

A firm that sends thirty pieces of correspondence a day gets back real time here. It gets consistency across staff as a side effect.

5. Answering questions from the matter instead of scrolling

"What's the settlement date?" "Has the deposit been paid?" "Did we tell the client about the easement, and when?"

Each one is a two-minute search through folders, emails and file notes, twenty times a day. Archie answers those from the matter itself.

It is the least glamorous item on this list and, in triSearch.ai's experience, the one practitioners notice first, because the interruptions stop.

Where AI does not save time, and shouldn't

The same honesty applies in the other direction. AI does not save time on the following.

Legal judgement. Whether a covenant matters, whether a special condition is acceptable, whether to advise a client to proceed. Those decisions carry liability and they stay with the licensed conveyancer or solicitor.

Advice. The output of a review tool is an input to your advice, not the advice.

Elapsed time on third parties. A council certificate takes as long as the council takes.

Novel or messy documents. Confident errors cluster where the input is unusual.

Regulators have said the same thing in stronger terms. According to the Law Society of NSW, legal profession regulators in NSW, Victoria and Western Australia have issued a joint statement on the responsible use of AI in legal practice, grounded in practitioners' existing conduct rules and duties.

In Queensland, Supreme Court Practice Direction 5 of 2025 requires that an individual practitioner be named as the responsible person for written submissions. According to the direction, practitioners responsible for submissions referring to non-existent cases or material may be referred to the Legal Services Commissioner or required to show cause why a costs order should not be made against them personally.

The direction applies to litigation, but its logic reaches every document a practitioner puts a name to. triSearch.ai's view is that a tool claiming to save you time on judgement, advice or third-party waits is a red flag, not a feature.

What has to be true for the savings to stick

Time saved on a task is only banked if three conditions hold.

You verify everything. Treat AI output as a first pass from a capable junior.

According to Lawyers Weekly's report on the VLSB+C Risk Outlook 2026, the Victorian regulator will treat failures to verify AI-generated material as extremely serious. The same report names AI and the new AML/CTF obligations as two of the three risk areas where lawyers need to build competency this year.

You use approved tools with the data inside the matter. According to Thomson Reuters' 2026 report, 34% of professionals use AI their organisation hasn't approved, and 41% lack access to tools that meet professional accountability standards.

A conveyancing file contains names, addresses, purchase prices, mortgage details and identity documents. Pasting that into a consumer chatbot is a data-handling problem before it is a productivity question.

Your AI runs where your matter runs. Savings evaporate if you have to copy output between systems.

triSearch.ai made that design choice from the start: Archie, Intake Forms and AI Contract Review sit inside triConvey and read from the matter. All triSearch AI features operate in a ring-fenced environment on Amazon Web Services, and client data is not shared outside triSearch or used to train AI models.

Swipe sideways to see the full table →

According to the Thomson Reuters report, 66% of professionals say AI meets or exceeds expectations where a clear AI strategy is in place, against 22% where there is none. For a conveyancing practice, a "strategy" doesn't need to be a document.

It needs to be a list of tasks, a tool the firm has approved, and a rule that nothing leaves the office unchecked.

How to tell whether your AI is actually paying off

Most firms don't measure. Only 18% of organisations track AI return on investment, according to Thomson Reuters, and 40% of professionals don't know whether it is measured at all. Doing the following for a fortnight puts a firm ahead of most of the profession.

Pick three tasks from the list above. Time them for a week without AI and a week with it. Count minutes per matter, not impressions.

Track rework. How often does the AI output need a substantive correction, as opposed to a tone edit? If the number is high on a document type, stop using it there.

Check the compliance path. Does intake data reach the AML/CTF record without being typed a second time? If not, the intake saving is smaller than it looks.

Ask staff which interruptions stopped. Matter Q&A pays off there, and it rarely shows up in a timesheet.

triSearch.ai would rather a firm run that test and find one task where the tool doesn't help than take a productivity claim on trust. AI in conveyancing in 2026 is a reading and drafting assistant with excellent recall and no judgement.

triSearch.ai's position is that every AI draft, flag and summary in a conveyancing file should be checked against the source document before it leaves the office, and that firms should measure the saving on named tasks from the first fortnight of use.