Voice Preservation & Procedural Fairness Agent for Melbourne Boutique Firms: Keep the Client’s Voice (and Your Disclosure Posture) Intact

You run a six-lawyer firm in Melbourne. A migration matter, an NDIS review, a Centrelink appeal — three different ART files this week and one paralegal. You’ve started using a model to help shape the first draft of statements and submissions. The problem isn’t speed; it’s that the third paragraph of your client’s statement now sounds like every other LLM output in the country, and you’re not sure whether the Tribunal expects you to say so. The Voice Preservation & Procedural Fairness Agent is built for that exact gap: keeping the client’s actual voice on the page, and keeping your AI-use disclosure posture defensible.

The problem

The Administrative Review Tribunal publishes practice directions and guidance for practitioners appearing before it, including expectations around expert evidence and the conduct of proceedings. Where AI is used to assist in preparing material filed with the Tribunal, the practitioner remains responsible for the accuracy, authenticity and fairness of what’s filed — and for disclosure where the framework requires it. Two failure modes follow from this for small firms:

Voice drift. A client’s statement, after a model has “cleaned it up”, reads in standard model prose. The applicant’s own cadence, vocabulary, and idiom — often the most probative material in a credibility-sensitive jurisdiction — is gone. On review, the Tribunal sees a generic statement that doesn’t sound like the person sitting in front of them.
Disclosure gaps. AI-use disclosure obligations vary by forum and matter type. A boutique firm running parallel ART, Federal Court, and state tribunal work doesn’t have a compliance team mapping which document needs which disclosure. The default of saying nothing is the risky default.

Both are compounded when one or two senior practitioners are reviewing everything the junior staff and paralegals produce.

What the Voice Preservation & Procedural Fairness Agent does

The agent sits between an AI-assisted draft and the version you file. It preserves the client’s voice when AI assists in drafting submissions and statements, and it surfaces the procedural fairness and disclosure decisions you need to make before lodgement — rather than after a directions hearing.

Concretely, it does three things:

Voice-diff against source material. It compares the AI-assisted draft against the raw client input (interview transcripts, file notes, the client’s own written account) and flags passages where the draft has departed from the client’s vocabulary, register, or factual emphasis.
Disclosure-trigger check. It tags sections of the document that were materially shaped by AI assistance and produces a draft disclosure statement aligned to the forum (ART, Federal Court, state tribunal) the matter is filed in.
Procedural fairness checkpoints. It identifies points in the submission where opposing material has been characterised, summarised, or rebutted — the places where fairness to other parties (and to the decision-maker) is most likely to be compromised by model-generated phrasing.

How it works

Upload the AI-assisted draft and the source material (client interview transcript, file notes, prior correspondence) into the agent.
Voice analysis runs locally against the source corpus, producing a per-paragraph similarity and divergence score relative to the client’s actual language.
Disclosure tagging identifies AI-assisted passages and maps them against the forum’s disclosure framework (ART Practice Directions, Federal Court GPN-AI, applicable state tribunal rules).
The agent returns a structured report: paragraphs to rewrite in the client’s voice, a draft AI-use disclosure paragraph for the relevant forum, and a checklist of procedural fairness checkpoints to review before filing.
The supervising solicitor signs off, edits the draft disclosure to reflect the firm’s policy, and lodges with an audit trail showing what the agent flagged and how it was resolved.

Why this matters in Melbourne

Melbourne boutique firms carry a disproportionate share of administrative review work — migration, NDIS, social security, veterans’ entitlements — where the credibility of an applicant’s account is often the central question. A statement that no longer sounds like the applicant is a statement that’s harder to defend on cross-examination and harder for a Tribunal member to credit. At the same time, Victorian solicitors are bound by the Australian Solicitors’ Conduct Rules, including the duty of candour to the court (Rule 19), which extends to AI-assisted material. For a small firm without a dedicated risk or compliance function, an agent that handles voice preservation and disclosure posture in the same pass is the practical way to keep both the client’s case and the firm’s professional standing intact.

Sources

Administrative Review Tribunal — Practice Directions and Other Guidance: https://www.art.gov.au/help-and-resources/professionals-and-practitioners/practice-directions-and-other-guidance
Federal Court of Australia — Use of Generative Artificial Intelligence Practice Note (GPN-AI): https://www.fedcourt.gov.au/law-and-practice/practice-documents/practice-notes/gpn-ai
Law Council of Australia — Australian Solicitors’ Conduct Rules: https://lawcouncil.au/policy-agenda/regulation-of-the-profession-and-ethics/australian-solicitors-conduct-rules

Exegesis capability references:

Join the waitlist

Join the waitlist — be first to know when access opens for Melbourne boutique firms

The Voice Preservation & Procedural Fairness Agent is in build. We’re scoping pricing (per-matter, per-user monthly, or firm-licence) based on what boutique firms tell us actually fits a small-team budget. Join the waitlist and the structure of the tier you sit in will be shaped by what we hear from firms like yours.