Privilege Perimeter Enforcement Gateway for Perth Boutique Firms: Stop Privileged Content Leaking Through Your AI Stack

Your firm is eight lawyers, two paralegals, and a shared workspace where everyone’s matters sit next to everyone else’s. Someone in litigation pastes a witness statement into a chatbot to summarise it. The summary is fine. The model retained the input. A week later, the same model — or its embeddings, or a fine-tune sitting on a vendor’s account — surfaces phrasing in a separate matter for a separate client. You have no audit trail, no way to prove the perimeter held, and an Administrative Review Tribunal expert evidence question coming up where you’d like to use AI tooling without explaining how privilege survived the round trip. The Privilege Perimeter Enforcement Gateway is built to make that boundary enforceable rather than aspirational.

The problem

Boutique firms run lean. The same lawyer drafting a Tribunal submission on Monday may be reviewing a commercial contract on Tuesday and instructing an expert witness on Wednesday. When AI tooling is introduced into that workflow — drafting assistants, summarisers, intake bots — the inputs and outputs cross matters in ways that traditional conflict checks and chinese walls weren’t designed for. The risks compound:

• Privileged material from Matter A becomes context for a prompt in Matter B
• Vendor terms permit retention, logging, or training on inputs unless explicitly disabled
• The Administrative Review Tribunal’s practice directions on expert evidence require transparency about how expert opinions are formed, and AI-assisted preparation is increasingly within scope
• The Australian Solicitors’ Conduct Rules require solicitors to maintain client confidentiality (Rule 9) and avoid conflicts (Rules 10–12) — duties that don’t pause when an LLM is in the pipeline
• The Privacy Act 1988 and OAIC guidance on AI impose obligations on how personal information is handled when AI systems process it

For a firm without an in-house security engineer, the practical question is: how do you know which prompts contained privileged content, where they went, and what came back?

What the Privilege Perimeter Enforcement Gateway does

The Gateway is a boundary control layer that sits between your team and any AI tool — internal or vendor-hosted — and enforces matter-level segregation on both inputs and outputs. Concretely, it:

• Classifies prompts and attachments by matter, client, and privilege status before they reach a model
• Blocks or redacts content that crosses defined perimeters (e.g. Matter A material cannot appear in a Matter B prompt context)
• Tags every output with its originating matter so it cannot be silently reused elsewhere
• Logs every crossing decision — allowed, blocked, redacted — to a tamper-evident audit trail
• Applies vendor-specific policies (no retention, no training, regional routing) at the gateway rather than relying on per-user settings

The deliverable is a deployed gateway, a written perimeter policy aligned to your matter taxonomy, and an audit log your principal can show a Tribunal or a professional standards review.

How it works

1. Perimeter mapping. We sit with your principal and practice manager and map the matter taxonomy, client groupings, and known conflict-sensitive boundaries. This becomes the policy file the gateway enforces.
2. Gateway deployment. The gateway is installed in front of the AI tooling your firm already uses — drafting assistants, summarisers, intake forms. Lawyers see no change in interface; the enforcement is upstream.
3. Classification and routing. Each prompt is classified against the perimeter policy. Privileged content is tagged, routed only to approved endpoints, and stripped of cross-matter context.
4. Audit logging. Every decision the gateway makes is logged with timestamp, user, matter, classification, and action. Logs are retained on-firm, not on a vendor’s account.
5. Review cycle. A monthly readout shows blocked crossings, near-misses, and policy gaps so the perimeter tightens over time rather than calcifying.

Why this matters in Perth

Perth boutique firms often handle a wider range of work per lawyer than their east-coast counterparts — resources and mining disputes, family law, commercial litigation, and administrative review work can sit inside the same practice. When matters span the Administrative Review Tribunal, the Federal Court, and the Supreme Court of Western Australia, the privilege and confidentiality obligations are layered, and the consequences of bleed aren’t theoretical. The ART’s practice directions on expert evidence expect transparency about how opinions and submissions are prepared. The Australian Solicitors’ Conduct Rules apply uniformly. The OAIC’s guidance on AI and privacy obligations applies regardless of firm size. A boundary you can demonstrate is a boundary you can defend.

Sources

1. Administrative Review Tribunal — Practice Directions and Other Guidance: https://www.art.gov.au/help-and-resources/professionals-and-practitioners/practice-directions-and-other-guidance
2. Law Council of Australia — Australian Solicitors’ Conduct Rules: https://lawcouncil.au/policy-agenda/regulation-of-the-profession-and-ethics/australian-solicitors-conduct-rules
3. Office of the Australian Information Commissioner — Privacy guidance on artificial intelligence: https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/hands-on-guides/guidance-on-privacy-and-the-use-of-commercially-available-ai-products

Exegesis capability references:

• Privilege Perimeter Enforcement Gateway spec
• RuleCheck by Exegesis

Join the waitlist

Join the waitlist — be the first to know when the Privilege Perimeter Enforcement Gateway opens to Perth boutique firms

We’re sizing the Gateway for firms under ten lawyers — the configuration, the policy templates, and the deployment effort all change at that scale. Join the waitlist and tell us about your matter mix; what we hear shapes how the boutique tier is built.