Barristers spend hours formatting citations they already know. The product is the document, not the database.
Before every hearing, barristers build a List of Authorities: every case, legislation, and book they intend to cite, formatted to court-specific templates. This process is entirely manual. Different courts require different structures. AGLC4 formatting rules are strict. Mistakes delay filings.
Built alongside Michael Green (practising barrister and primary stakeholder) and Sean Simpson (librarian SME) at BarNet OpenLaw.
Onboarding: search and join your chambers to access the shared library
Mapping the system before the screens
I worked through the full information architecture on paper before moving to Figma. Two nav structures (barrister 5 items, clerk 7 items), two daily flows (Find, Cite, Track for barrister; Track, Scan, Manage for clerk), and the points where they intersect.
Two roles, two navigation structures, and where they meet. Barrister and clerk use the same product but live in different halves of it.
Five cross-role interactions that make the product work as a single system. If IX-5 (shared catalogue) doesn't work, nothing else does.
A phased product, not a feature list
I structured the product as three phases, not a single roadmap. Phase 1 ships without AI: manual research, authority list building, AGLC export, library management. Phase 2 layers AI on top only after Phase 1 adoption is validated. Phase 3 is the strategic direction: matter workspace and portable brief, competing with eBriefReady.
The phased structure forced a hard question for each feature: what's the minimum that makes this useful, and what does it enable next?
Three phases, three different product bets. Each phase ships independently.
Dashboard: active authority lists, research blockers, and alerts surfaced on login
Authority list as the core product, not the library
Early framing positioned this as a "library management system." But the real value is the authority list: the court-ready document barristers file before every hearing. The library is a supporting feature for clerks. The authority list is the product.
Barristers search cases via JADE, add items instantly, and manually add uncatalogued materials. Manual input is not an edge case. It's a core system input that feeds into the clerk's cataloguing workflow.
Authority list editor (left) with live court-formatted preview (right). VIC Supreme Court template with AGLC4 citation formatting.
AGLC compliance from court templates, not from a style guide
I researched the actual filing templates from four courts: VIC Supreme Court (Court of Appeal), NSW Supreme Court (Court of Appeal), Federal Court (GPN-AUTH), and the High Court of Australia. Each has different requirements.
"List of Authorities" vs "Table of Authorities." Part A/B/C structures with court-specific wording. Continuous numbering across parts. Party name formatting. AGLC4 italicisation rules for case names. These details determine whether a filing is accepted or returned.
Uncatalogued books as structured input
Incomplete citations, missing pinpoints, and uncatalogued materials are flagged in the editor with status indicators. Errors are surfaced where they can be acted on, not buried in the document output.
This creates a natural flow: barristers resolve what they can, and unresolved items surface as signals to the clerk's catalogue workflow.
Editor flags incomplete citations and missing pinpoints with inline status indicators
Chambers library: shared catalogue with availability tracking and loan management (clerk-facing)
Chambers Library
The library is the clerk's domain: catalogue management, loan tracking, availability, and acquisition. Barristers can browse and add items directly to their authority lists.
Supporting system, not the core product. But it's how the clerk's work connects to the barrister's workflow, and how uncatalogued materials get resolved over time.
List privacy as foundation, not feature
Barristers in the same chambers can represent opposing sides of the same matter. An authority list reveals legal strategy. If one barrister can see another's list for the same case, the system creates a conflict of interest.
List privacy is non-negotiable from day one. It's an architectural constraint that shapes data access, sharing permissions, and how the chambers library interacts with individual lists. Identifying this early prevented a structural redesign later.
AI with a hard constraint
Phase 2 introduces AI authority suggestion, adverse case detection, and pinpoint inference. Every design decision is downstream of one rule: the AI must not invent authorities. Rule 24 of the Barristers' Rules 2015 means a barrister who submits a non-existent case to court has committed professional misconduct. Retrieval from JADE index only, no LLM citation generation, every suggestion links to source.
The spec also captures a strategic decision point: add CaseTrace (verified paragraph-level pinpoints) to the JADE integration, or ship without it? The decision shapes both feature quality and competitive moat, but requires internal BarNet alignment.
Rule 24 of the Barristers' Rules 2015: a barrister who submits a non-existent case to court has committed professional misconduct. The AI layer is designed around this single constraint.
Portable Brief direction
Mid-project, the barrister stakeholder surfaced a bigger vision: extend CLMS into the eBriefReady competitor space. eBriefReady owns the delivery layer (solicitor to barrister brief transfer, bundling, OCR search, AI chronology) but has no research layer. Our differentiation is the piece they're missing: JADE-linked legal research, authority list building, AGLC auto-formatting.
The architectural question stays open: is this an extension of CLMS (one product, matter workspace as a new module) or a separate product sharing the same design system? Phase 1 adoption is the gate. If barristers aren't using authority lists, the matter workspace has no foundation.
Phase 3 direction: extending into the eBriefReady space with matter workspace and portable brief.
The depth is in the domain.
This project required understanding how Australian courts actually work: which templates they accept, how citations must be formatted, what happens when an authority is uncatalogued, and why a barrister's list is strategically sensitive.
The system design, court template research, and Phase 1 architecture were developed through close collaboration with Michael Green (barrister) and Sean Simpson (librarian), with multiple rounds of review. Phase 2 (AI) and Phase 3 (Portable Brief / eBriefReady) are actively in progress.
