Harvey Ai, Lawlux & Liam Barnes - Legal Tech Blog - October 2026
Emerging technologies are going to shape your legal career, so you need to know how they work before you meet them on the job. This blog series shows how new tools are used to solve everyday legal queries. Where possible, we include a short video so you can see it for yourself.
—
Meet Harvey AI - The US$15 billion legal tech company you need to know about.
Harvey AI is an AI platform built for legal work. In its demo, a Harvey product manager uploads a treaty and asks for a client memo summarising it.
Legal Question - Who taxes the profits when an Australian mining company earns money through a South African subsidiary?
Cross-border tax is a staple of commercial law, and the starting problem is simple. If an Australian mining company earns profits through a subsidiary in South Africa, both countries can claim the right to tax the same income. A tax treaty, or double tax agreement, is what decides which country taxes what and at what rate. When a treaty changes, the parent company needs to know what that means for its tax bill, and its lawyer needs to explain it in plain terms.
How does Harvey help a lawyer answer this legal question?
Harvey reads the document and drafts the memo in an editor with a chat panel alongside. She asks for a missing section, “tax residency scenarios”, and Harvey adds it as a new version, with the changes shown as a tracked redline. She can then check the citations, compare earlier versions and edit by hand. (The video’s example is a UK-Cyprus treaty but the workflow is the same.)
For a student or junior paralegal, the lesson is the workflow. The first draft arrives in minutes, but you still decide what’s missing, check each point against the treaty, and own the final advice. The redline and version history are what make that review practical.
Meet Lawlux - An Australian employment law firm that has built its own AI workspace to answer employers’ questions from decided cases.
Legal Question - Can you dismiss a manager for drinking on their lunch break?
Employment law students meet this one early, and “it depends” isn’t an answer. Under the Fair Work Act 2009 (Cth), a dismissal is unfair if it was harsh, unjust or unreasonable, and whether there was a valid reason is a key consideration. A manager drinking during a lunch break raises a familiar problem: how far does the employer’s policy reach, and what standard applies to someone in a leadership role?
Lawlux built Reassure, a client AI workspace for employers and people-and-culture teams, to work through questions like this. In its demo, the question is typed in plain language. Reassure searches Lawlux’s database of 77,000 employment law decisions (updated daily, according to Lawlux), plans the research, and in the demo draws on 62 sources. Rather than hedging, it matches the facts to similar decided cases. In the example shown, a case involving a store manager drinking on a lunch break turned on the higher standard expected of managers. Each source comes with an extract and a link through to the full decision.
For a student or junior paralegal, the lesson is that good research means finding the case whose facts sit closest to yours. Reassure does the first pass, and you test whether the facts really line up. When the matter needs more, such as a show cause letter here, the user can escalate to a Lawlux lawyer with the whole conversation attached.
How does Reassure help an employment lawyer or their client answer legal questions?
See David Burnton discuss Reassure in this recent Linkedin post.
Five ways to get more value from AI tools
Liam Barnes is a consultant who trains legal teams on AI, walks through five Claude features for legal work. These habits apply to any AI tool and show you that the true value from AI comes from the input as much as the output!
- Structure your prompt. Give the tool five things: the context (who you’re acting for), the task, the standards to measure against, the tone (who it’s written for) and the format, such as a table with red, yellow and green risk ratings. The same lease goes from a wall of text to a partner-ready table.
- Keep matter context in one place. Set up a workspace per matter with the deal details and instructions saved, so you don’t re-explain the matter in every conversation.
- Write your firm’s playbook down once. A plain-text file of standard positions, for example how the firm treats non-solicitation or governing law, can be applied automatically to every review.
- Give it a job, not a question. Ask for a finished deliverable, such as a comparison of three contracts in a spreadsheet, and check the plan before it runs.
- Use ready-made legal commands. Built-in commands can review a contract against your playbook or triage an NDA.
Yegal is Australia's first on-demand paralegal service. We connect law firms and in-house legal teams with contract paralegals for short, medium and long-term placements. Requesting your next paralegal is quick and easy. CLICK HERE to get started.
