# LEAP Marketplace Plugin — Product Brief

**Prepared by:** Mark Harley, Principal Solicitor, Boss Lawyers Pty Ltd
**Date:** May 2026
**Status:** Confidential — Draft v1.0
**Purpose:** Commercial product brief for development, investment, and partnership discussions

---

## Executive Summary

Australian litigation firms run their practice on LEAP but think with paper, instinct, and tribal knowledge. A junior solicitor at a $400/hr firm and a principal at a $1,200/hr firm both manage cases the same way: file notes, mental checklists, and the hope that nothing falls through. The frameworks that separate good litigators from bad — element-evidence mapping, gap analysis, merit/risk/commercial scoring, costs-defensible billing — exist only in the heads of senior practitioners.

This product productises those frameworks. It is a LEAP Marketplace plugin that adds an AI litigation intelligence layer over existing LEAP matter data. It does not replace LEAP. It makes LEAP firms litigate better.

The product is built by a practising commercial litigation solicitor, not a tech company. That is the moat.

**Target ARR:** $4.2M by end of Year 3, 70–80% gross margin.
**MVP build cost:** $80–150k outsourced, or $40–80k with AI-assisted in-house build.
**Key risk:** LEAP API/marketplace dependency.

---

## 1. Product Vision

### 1.1 Product Name Options

Five candidates, each tested for: (a) availability of `.com.au` and `.com`, (b) trademark conflict on IP Australia search, (c) legal-market fit, (d) memorability.

| Rank | Name | Reasoning |
|------|------|-----------|
| 1 | **LitIQ** | Short, sharp, two syllables. Conveys litigation + intelligence. Easy to say in conferences. Strong domain availability prospects. Reads as a noun ("run it through LitIQ") and a brand. |
| 2 | **CaseRank** | Descriptive, immediately understood. Risk: too descriptive — harder to trademark distinctively. Better as a feature name within a broader product. |
| 3 | **Counsel** | Premium feel, but generic and likely conflicts with Counsel.ai and existing legal services. Skip. |
| 4 | **Merit** | Clean single-word brand. Refers to the central concept (case merit). Strong, but the word is heavily used in the legal vernacular and unlikely to be available as a TM in Class 9/42 for legal software. |
| 5 | **Argus** | Mythological reference (the all-seeing). Premium, distinctive, defensible TM. Risk: less self-evident; needs marketing investment. |
| 6 | **BossAI / Boss Intelligence** | Ties to Boss Lawyers brand. Risk: limits sale/spin-out value, conflates legal practice with software, and "Boss" is overloaded in tech (BossDB, BossNote, etc.). Avoid as primary. |
| 7 | **Praecedo** | Latin "I go before / I lead." Premium feel, defensible. Risk: pretentious for SME market. |

**Recommendation:** **LitIQ** as primary brand. CaseRank, Element Tracker, Gap Detector remain as feature names inside the product.

### 1.2 One-Line Pitch

> **"AI litigation intelligence for LEAP firms."**

Alternate longer pitch:
> "LitIQ is the AI litigation intelligence layer for Australian law firms running LEAP. It scores your matters for merit, risk and commercial viability; tracks evidence against the elements of every cause of action; flags the gaps before opposing counsel finds them; and drafts costs-assessment-defensible billing narratives — all from your existing LEAP data."

### 1.3 The Problem

Australian SME litigation firms (1–50 fee earners) face a structural intelligence gap:

1. **No element-evidence mapping.** Causes of action are pleaded, then everyone hopes the evidence covers each element by trial. Most firms have no formal way to map evidence to elements.
2. **No matter scoring.** Senior practitioners "know" which matters are strong and which are dogs. Junior practitioners don't, and the seniors don't have time to triage every matter weekly.
3. **No structured gap analysis.** Procedural deadlines, missing witness statements, unaddressed defences — these are caught reactively, often when prejudicial.
4. **Billing narratives are a tax on profitability.** Solicitors either under-narrate (write-offs on assessment) or over-spend partner time fixing junior narratives. The work has been done; the words to describe it haven't.
5. **Knowledge banks are folder structures, not searchable intelligence.** Precedent libraries are dead documents. There is no semantic search across the firm's authorities, prior pleadings, or successful submissions.
6. **Document drafting is high-leverage but error-prone.** AI tools exist but are generic, hallucinate authorities, and don't integrate with matter data.

### 1.4 The Solution

A single LEAP Marketplace plugin that:

- Reads matter data from LEAP (parties, documents, time entries, calendar, correspondence).
- Adds an intelligence layer: scoring, mapping, tracking, drafting, retrieving.
- Pushes back to LEAP only what belongs there: time entries, generated documents, calendar reminders.
- Keeps AI outputs in the plugin where they can be iterated, reviewed, and audited.

The plugin is **not** a practice management system. It is a **litigation intelligence overlay** on LEAP.

### 1.5 Target Market

**Primary (Year 1):**
- Queensland commercial litigation and insolvency firms
- 1–20 fee earners
- Already on LEAP
- Doing $1M–$15M revenue
- Principal-led, commercially conservative

**Secondary (Year 2+):**
- NSW and Victorian litigation firms
- Federal Court practitioners (Corporations List, IP, competition)
- Boutique insolvency practices
- Mid-tier firms (20–80 fee earners) on LEAP

**Tertiary (Year 3+):**
- Plaintiff personal injury firms (different evidence model — partial fit)
- Family law litigation (limited fit; evidence model differs)
- New Zealand market (LEAP has presence)

### 1.6 Differentiator

| Generic Legal AI | LitIQ |
|------------------|-------|
| Built by tech founders, validated by lawyer advisors | Built by a practising commercial litigator who uses it daily |
| Generic prompts, US-trained models, US authorities | QLD-first, Australian authorities, UCPR-aware |
| Stand-alone web app, manual data entry | LEAP-native, zero-duplication of matter data |
| One-size-fits-all features | Litigation-specific, opinionated, framework-driven |
| "AI for law" marketing | "Element tracking, gap detection, costs-defensible billing" — concrete |

The moat is **dogfooding**. Every framework in the product is in active use at Boss Lawyers. Every feature ships against real matters before it ships to customers.

---

## 2. Market Analysis

### 2.1 LEAP Market Position

LEAP Legal Software is the dominant SME practice management platform in Australia:

- **Estimated installed base:** 7,000–9,000 Australian law firms (publicly stated "over 100,000 lawyers globally" across LEAP, ALB, Linetime brands)
- **Market share:** ~50–60% of SME law firm market in AU (1–50 fee earners)
- **Marketplace:** LEAP App Marketplace exists with ~30–60 third-party integrations
- **API:** Public LEAP Public API (REST) with OAuth2; documented endpoints for matters, contacts, documents, time, calendar

### 2.2 Total Addressable Market (Australia)

| Segment | Firms | LitIQ-Eligible | Notes |
|---------|-------|----------------|-------|
| QLD litigation firms on LEAP | ~600 | ~400 | Initial beachhead |
| NSW litigation firms on LEAP | ~1,200 | ~800 | Year 2 expansion |
| VIC litigation firms on LEAP | ~900 | ~600 | Year 2 expansion |
| Other states/territories | ~700 | ~400 | Year 3 |
| **Total AU TAM** | **~3,400** | **~2,200** | At avg 4 users/firm = **~8,800 seats** |

At blended ARPU of $300/user/month, full TAM = **~$31.7M ARR**. Realistic 5-year ceiling at 25% market penetration = **~$8M ARR**.

### 2.3 Competitive Landscape

**On LEAP Marketplace (as of May 2026):**
- *Document automation*: HotDocs, Lexis Affinity, Smarter Drafter — template-based, not AI-driven
- *eSignature*: DocuSign, Adobe Sign — utility integrations
- *AI assistants*: emerging but generic; no LEAP-native litigation intelligence product currently dominates
- *Time tracking*: WiseTime, Smokeball Auto Time — adjacent, not competitive

**Off-marketplace AI legal tools competing for budget:**
- *Harvey* — enterprise (top-tier firms), $100k+ ACV, not SME
- *Spellbook* — contract drafting, not litigation
- *CoCounsel (Thomson Reuters)* — research-led, not LEAP-integrated
- *Lexis+ AI* — research, not litigation workflow
- *Josef* — automation/intake, not intelligence
- *Lawpath / LegalVision tools* — consumer-end, not firm-grade

**Key finding:** No incumbent owns "AI litigation intelligence for LEAP firms." Window open ~12–24 months before Harvey/CoCounsel push down-market or LEAP builds in-house.

### 2.4 Pricing Benchmarks (AU and global, AUD)

| Product | Price/user/month | Segment |
|---------|-------------------|---------|
| Harvey | ~$300–500 | Top-tier |
| CoCounsel | ~$100–150 | Mid-tier |
| Spellbook | ~$120 | Contract |
| Lexis+ AI | ~$200–300 | Research add-on |
| Smokeball | ~$120–199 (full PMS) | SME PMS |
| LEAP | ~$140–180 (full PMS) | SME PMS |
| Clio Manage + Duo | ~$200 | SME PMS |

Sweet spot for an add-on plugin: **$149–$499/user/month** (must be priced below LEAP itself for the lower tier; can exceed LEAP at the enterprise tier where the value is documented).

### 2.5 Why Litigation Specifically

- Highest hourly rates → highest willingness to pay
- Highest write-offs on costs assessment → highest measurable ROI
- Most complex evidence/element mapping → most automation value
- Highest stakes per matter → most willingness to adopt a "second pair of eyes"
- Senior partner time is the binding constraint → tools that leverage juniors yield direct margin

Estimated per-firm value at a 5-fee-earner litigation firm: **30–60 hours/year saved on billing narratives**, **5–15% reduction in costs-assessment write-offs**, **1–2 strategically improved matter outcomes per year**. At $800/hr, that is $24–$48k/year of recovered value against ~$18k/year in licence cost. ROI 1.3x–2.7x conservative.

### 2.6 Beachhead Strategy

**QLD-first** because:
- Mark practises here; regulatory familiarity is deep (UCPR, BIF Act, QCAT)
- QLD Law Society network is accessible
- QSC/QDC procedural quirks are most automatable when you live them
- Smaller than NSW/VIC means achievable share before national push

---

## 3. Product Features (MVP — v1.0)

The MVP rests on four internal frameworks already in development at Boss Lawyers:

- `reference/case-ranking-framework.md` — merit/risk/commercial scoring rubric
- `reference/evidence-assessment-framework.md` — 5-dimension evidence scoring
- `reference/knowledge-bank-architecture.md` — semantic search and authority retrieval
- `reference/document-generation-spec.md` — drafting pipeline with quality gates

### 3.1 v1.0 — MVP (Months 1–6)

#### Feature 1: Case Ranking Dashboard
- Pulls matters from LEAP via API
- Solicitor inputs key matter facts (cause(s) of action, parties, quantum, key dates)
- Plugin generates three scores (0–100) per matter:
  - **Merit Score** — likelihood of success on the law and pleaded facts
  - **Risk Score** — adverse costs, counterclaim, reputational, recovery risk
  - **Commercial Score** — quantum vs cost-to-judgment, payability, time-to-resolution
- Dashboard view: all matters, sortable, with traffic-light visualisation
- Drill-down view: per-matter breakdown of each score with reasoning
- Re-scoring on demand (after key event: pleading filed, evidence served, etc.)

#### Feature 2: Element Tracker
- For each matter, user selects causes of action from a curated taxonomy (breach of contract, misleading and deceptive conduct s 18 ACL, ss 180–184 Corporations Act duties, oppression s 232, voidable transaction Pt 5.7B, etc.)
- Plugin loads the elements of each cause from the knowledge bank
- User maps available evidence to each element
- Visualisation: matrix of elements (rows) × evidence (columns) with coverage indicators
- "Elements at risk" alert when any element has zero or weak evidence

#### Feature 3: Evidence Inventory
- Catalogue of all evidence linked to a matter (documents pulled from LEAP, statements drafted, expert reports, real evidence)
- Each evidence item scored on five dimensions (per `reference/evidence-assessment-framework.md`):
  - **Relevance** — to which element(s)
  - **Reliability** — source credibility, contemporaneity, corroboration
  - **Admissibility** — Evidence Act 1977 (Qld) / 1995 (Cth) issues, hearsay, opinion, privilege
  - **Weight** — likely persuasive force at trial
  - **Availability** — secured, in possession, requires subpoena, witness availability
- Aggregate evidence score per element drives the Element Tracker visualisation

#### Feature 4: Gap Detector
- Three-layer gap analysis:
  - **Substantive gaps** — elements with no/weak evidence
  - **Procedural gaps** — UCPR-driven (no defence filed, no disclosure, no list of documents, no expert directions made), surfaced from LEAP calendar/document data
  - **Strategic gaps** — anticipated defences not addressed, limitation issues, jurisdiction issues
- Each gap has a recommended action ("draft request for further and better particulars", "issue subpoena to non-party X", "file r 444 letter")
- Daily digest email per user, weekly digest per principal

#### Feature 5: Document Drafter (Basic)
- Initial document types (QLD-first):
  - Letters of demand
  - Statutory demands (s 459E Corporations Act compliant)
  - Form 17 defences (per Boss Lawyers' QLD pleading guide)
  - r 444 / r 445 letters
  - Originating applications and supporting affidavits (skeletal)
  - Costs-assessment-ready bills of costs (narrative only; not Form 60 yet)
- Drafts use matter data from LEAP and reasoning from the Element Tracker
- Quality gates (per `reference/document-generation-spec.md`):
  - Authority verification (no hallucinated cases)
  - UCPR rule citation check
  - Style guide compliance (Boss Lawyers QLD pleading guide as default; firm-customisable)
  - Mandatory human review checkpoint before export

#### Feature 6: Knowledge Bank (Starter)
- Pre-loaded with QLD-curated authorities by practice area:
  - Pleading authorities (*Dare v Pulham*, *Banque Commerciale v Akhil*, *Bruce v Odhams Press*, *Groves v ALHMWU*, *Chan v Goldenwater*)
  - Contract and ACL authorities
  - Corporations Act / insolvency authorities
  - Costs assessment authorities
- Semantic search via pgvector
- Firm-specific additions: users add their own authorities, prior pleadings, internal precedents
- Citation export in AGLC4 format

#### Feature 7: Billing Narrative Generator
- Reads time entries from LEAP
- Reads document and email activity from LEAP
- Generates costs-assessment-defensible narratives in conventional QLD phrasing per `TOOLS.md` rules
- Each entry: What + From/To Whom + Date + About What + Why
- User reviews, edits, approves, then pushes refined narratives back to LEAP time entry descriptions
- **This is the highest-ROI feature for adoption** — every solicitor feels this pain weekly

### 3.2 v2.0 Roadmap (Months 7–18)

- **Strategy Playbooks by matter type** — codified workflows for common matter types (oppression action, statutory demand defence, freezing order application, DOCA challenge)
- **Opponent Intelligence** — anonymised, firm-level statistics: "Firm X typically files defences within 28 days, settles 60% of disputes pre-trial, runs 12% to judgment". Built from public court data.
- **Witness Assessment Module** — score lay and expert witnesses on credibility, reliability, prior performance under cross-examination (where public)
- **Settlement Recommendation Engine** — Calderbank optimiser: given current evidence position and quantum spread, recommends an offer, timing, and structure
- **Multi-jurisdiction support** — NSW UCPR, VIC Civil Procedure Act/Rules, Federal Court Rules 2011, Federal Court (Corporations) Rules
- **Advanced document drafting with quality gates** — submissions, expert briefs, full pleadings with full r 150 / r 165 / r 166 compliance checks
- **Client reporting portal** — sanitised, client-facing matter status views (case score → "strong/moderate/at risk" lay translation)

### 3.3 Explicitly Out of Scope (Forever)

- Trust accounting (LEAP keeps it)
- Invoicing and statements (LEAP keeps it)
- Conflict checks (LEAP / dedicated tools)
- Court e-filing (handled by eCourts portal)
- Discovery review at scale (Relativity / Reveal territory)
- Contract review/negotiation (Spellbook territory)

---

## 4. LEAP Integration Architecture

### 4.1 LEAP Public API — Data Pulled

| Endpoint | Data | Use |
|----------|------|-----|
| `/matters` | Matter metadata, parties, status, area of law | Case ranking, element tracker |
| `/contacts` | Parties, opposing solicitors, counsel, witnesses | Element tracker, opponent intel |
| `/documents` | Matter documents (metadata + content where permitted) | Evidence inventory, gap detector |
| `/time-entries` | Time recorded per matter per fee earner | Billing narrative generator |
| `/calendar` / `/critical-dates` | Hearings, directions, deadlines | Procedural gap detector |
| `/correspondence` | Email metadata where surfaced via LEAP | Activity inference for billing narratives |
| `/users` | Fee earner identity, hourly rate, level | Billing narrative attribution |

### 4.2 Authentication

- **OAuth2** authorisation code flow via LEAP Marketplace
- Per-firm tenancy; no cross-firm data access
- Refresh tokens stored encrypted (AES-256) in plugin database
- Scoped permissions: read matter/contact/document; read/write time entry; read/write calendar; read user

### 4.3 Data Flow

```
┌─────────┐      OAuth2       ┌──────────┐
│  LEAP   │ ◄──────────────►  │  LitIQ   │
└─────────┘                   └──────────┘
     │                              │
     │  GET /matters                │
     │  GET /documents              │
     │  GET /time-entries           │
     │ ───────────────────────────► │
     │                              │  ┌───────────────┐
     │                              │  │   AI Layer    │
     │                              │ ─►│  - Scoring    │
     │                              │  │  - Mapping    │
     │                              │  │  - Drafting   │
     │                              │  └───────────────┘
     │                              │
     │  POST /time-entries          │
     │  POST /documents             │
     │  POST /calendar-events       │
     │ ◄─────────────────────────── │
```

### 4.4 What Stays in LEAP vs in LitIQ

| Stays in LEAP | Lives in LitIQ |
|----------------|-----------------|
| Trust accounting | Case scores |
| Invoicing | Element-evidence mapping |
| Source-of-truth matter records | Evidence assessments |
| Source-of-truth time entries | Gap analysis state |
| Source-of-truth documents | Drafting workspace (until exported) |
| Trust ledger | Knowledge bank (firm-specific additions) |
| Conflict checks | Strategy playbooks |
| Client/matter inception | AI prompts and outputs (auditable) |

### 4.5 Data Storage and Sovereignty

- **Region:** AWS Sydney (`ap-southeast-2`) — exclusively
- **No US, EU or APAC-other replication**
- **Encryption at rest:** AES-256 (RDS, S3, EBS)
- **Encryption in transit:** TLS 1.3 minimum
- **Backups:** Encrypted, retained 35 days, also in `ap-southeast-2`
- **AI inference:** Anthropic Claude via Sydney region endpoint where available, otherwise via private routing with no-retention agreement; fallback to on-prem inference for enterprise tier
- **No training on customer data** — contractual prohibition with all AI providers; documented in customer agreement

### 4.6 Privacy Posture

- Australian Privacy Principles (APPs) compliance documented in Privacy Policy
- DPA with each customer firm (joint controllership where applicable)
- Customer firm remains the data controller for client data; LitIQ is processor
- Customer-managed encryption keys (CMEK) available at Enterprise tier
- Right to delete on contract termination; export to JSON within 30 days

---

## 5. Technical Architecture

### 5.1 Stack

| Layer | Technology | Reasoning |
|-------|-------------|-----------|
| Frontend | Next.js (React 18, TypeScript) | Embeddable in LEAP iframe; standalone fallback; mature ecosystem |
| UI library | Radix + Tailwind | Accessible, lawyer-friendly density |
| Backend API | Node.js (NestJS) or Python (FastAPI) | NestJS preferred — TS shared types with frontend; FastAPI if Python ML-heavy |
| Database | PostgreSQL 16 + pgvector | Relational core + semantic search in one engine |
| Object storage | AWS S3 (Sydney) | Document/evidence storage |
| Queue/jobs | AWS SQS + worker (Node) | Async AI calls, batch scoring |
| AI | Anthropic Claude (Opus for drafting, Sonnet for retrieval/scoring) | Best Australian-context performance, best output discipline |
| Embeddings | Voyage AI / OpenAI text-embedding-3-large | Knowledge bank semantic search |
| Auth (firm) | LEAP OAuth2 | Required for marketplace |
| Auth (user) | Auth0 / Clerk + LEAP user mapping | Per-user permissions inside the firm |
| Hosting | AWS Sydney (`ap-southeast-2`) | Sovereignty mandate |
| Compute | ECS Fargate | Right-sized, no server management |
| CDN | CloudFront with `ap-southeast-2` origin | Performance |
| Observability | Datadog or Grafana Cloud (Sydney) | SLA monitoring |
| CI/CD | GitHub Actions + AWS CodeDeploy | Standard |
| IaC | Terraform | Reproducible environments |

### 5.2 Multi-Tenancy

- **Tenant isolation:** logical (single DB, tenant_id row-level security in Postgres) at Starter/Professional; physical (separate DB) at Enterprise
- **Compute:** shared, with per-tenant rate limits and quotas
- **AI calls:** per-tenant key isolation; per-tenant prompt logging for audit
- **Data export:** per-tenant JSON export at any time

### 5.3 Security

- SOC 2 Type II target by end of Year 2
- ISO 27001 alignment from Day 1
- Penetration test before public launch and annually
- Bug bounty program from Year 2
- All admin access via SSO + hardware MFA
- Audit log of every AI prompt and output retained 7 years (matches solicitor record-keeping obligations)

---

## 6. Business Model

### 6.1 Pricing

| Tier | Price (per user/month, ex GST) | Features |
|------|---------------------------------|----------|
| **Starter** | $149 | Case Ranking Dashboard, Element Tracker, Gap Detector, Billing Narrative Generator |
| **Professional** | $299 | Starter + Document Drafter, Knowledge Bank, Evidence Inventory |
| **Enterprise** | $499 | Professional + Strategy Playbooks, Opponent Intelligence, Settlement Engine, custom playbooks, CMEK, dedicated support, single-tenant DB |

- Annual prepay = 2 months free (effectively 16% discount)
- 14-day free trial on Starter/Professional
- Enterprise: minimum 5 seats, annual contract
- Implementation fee on Enterprise: $5k (data migration, custom playbook ingestion)

### 6.2 Revenue Projections

| | Year 1 | Year 2 | Year 3 |
|---|--------|--------|--------|
| Firms | 20 | 80 | 200 |
| Avg users/firm | 3 | 4 | 5 |
| Total seats | 60 | 320 | 1,000 |
| Blended ARPU/user/month | $299 | $299 | $349 |
| **ARR (year-end run rate)** | **$215k** | **$1.15M** | **$4.19M** |
| Mix (Starter/Pro/Ent) | 30/65/5 | 25/65/10 | 20/60/20 |

Assumptions:
- Year 1: 20 firms = ~5% of QLD TAM. Achievable with founder-led sales, beta-to-paid conversion, and LEAP Marketplace listing.
- Year 2: 80 firms = ~3% of national TAM. Requires NSW/VIC launch and 1 dedicated CSM.
- Year 3: 200 firms = ~6% of TAM. Requires content marketing, referral program, and at least one strategic partnership (chambers, association).

### 6.3 Cost Structure

| Cost (monthly, at scale) | Year 1 (avg) | Year 2 (avg) | Year 3 (avg) |
|---------------------------|--------------|--------------|--------------|
| AI API (Claude) | ~$300 | ~$2,500 | ~$10,000 |
| AWS hosting | ~$1,500 | ~$4,000 | ~$8,000 |
| SaaS tools (Auth, Datadog, etc.) | ~$500 | ~$1,500 | ~$3,000 |
| Founder time (notional) | $20,000 | $20,000 | $25,000 |
| Dev/contractor | $10,000 | $25,000 | $50,000 |
| Customer success | $0 | $9,000 (1 FTE from M18) | $15,000 |
| Sales/marketing | $1,000 | $5,000 | $20,000 |
| **Total monthly OPEX** | ~$33k | ~$67k | ~$131k |

### 6.4 Margins

- Variable cost per user (AI + hosting): ~$5–15/user/month
- Gross margin: 70–80% at Professional tier; 80–85% at Enterprise
- Contribution margin (after CSM allocation): 60–70% at scale
- LTV/CAC target: >3 by end of Year 2; >5 by Year 3

### 6.5 Capital Strategy

**Bootstrap path (recommended for Year 1):**
- Boss Lawyers funds MVP build ($80–150k) from operating cash
- Eligible for R&D Tax Incentive 43.5% refundable offset (group turnover < $20M) — recovers ~$35–65k of MVP cost
- Reach $215k ARR by end of Year 1 → cash-flow break-even on operating costs (excl. founder time)

**Capital raise option (Year 2 if growth accelerates):**
- Seed round: $1.5–2.5M at $8–12M pre-money
- Use of funds: 2 engineers, 1 CSM, 1 sales hire, NSW/VIC launch
- Investor profile: AU early-stage SaaS funds (Tidal, Black Sheep, Folklore), legal-tech-aware angels

**Cap table considerations:**
- Founder (Mark): 100% at outset
- Tech co-founder: 15–25% if needed (post-MVP would prefer not to dilute this much; pre-MVP this is the realistic price of velocity)
- ESOP pool: 10% from incorporation
- Seed dilution: 15–20% typical at $1.5–2.5M raise

---

## 7. Go-to-Market Strategy

### 7.1 Phase 1 — Build MVP (Months 1–3)

- Design partner: Boss Lawyers (active matters as test bed)
- 3–5 beta firms recruited from QLD litigation network (no cost; quarterly feedback commitment)
- Core features: Case Ranking, Element Tracker, Gap Detector, Billing Narrative Generator
- Brand identity, domain, IP Australia trademark application
- Privacy Policy, Terms of Service, DPA template

### 7.2 Phase 2 — QLD Launch (Months 3–6)

- LEAP Marketplace listing (review and approval cycle: 4–8 weeks; budget for it)
- QLD Law Society and QLS Symposium presence; sponsor a stream
- Content marketing series: "How AI is changing QLD litigation practice" — 12 articles, weekly cadence, by Mark
- Webinar series: monthly, recorded, gated (lead capture)
- Referral program: existing customer gets 1 month free per referred firm that signs up
- Case study: one Boss Lawyers matter end-to-end with quantified time/cost savings

### 7.3 Phase 3 — National Expansion (Months 6–12)

- NSW and VIC jurisdiction support (UCPR-NSW, Civil Procedure Act/Rules VIC)
- Federal Court support (FCR 2011, FC(C)R 2000) — high value because Corporations Act work runs Federal
- Partnership with at least one barristers' chambers for authority validation and credibility
- Podcast: "The Australian Litigator's Edge" — 26 episodes, weekly
- Targeted outbound to NSW/VIC LEAP firms (NSW Law Society Litigation Section, VIC LIV Civil Litigation Section)

### 7.4 Phase 4 — Scale (Year 2+)

- Enterprise tier and dedicated success
- White-label option for large firms (top-100 AU law firms; consider Macpherson Kelley, Mills Oakley as targets)
- API to integrate with **Smokeball, SILQ, Actionstep** — diversifies away from LEAP risk
- New Zealand market entry (LEAP has NZ presence; rules differ — High Court Rules 2016)
- Channel partnerships: legal accountants, practice management consultants

### 7.5 Marketing Channels (Realistic for AU Legal)

- LEAP Marketplace listing — primary discovery
- Mark's LinkedIn (highest ROI: practising solicitor with credible voice)
- QLS / NSW Law Society / LIV publications and events
- Australian Lawyer / Lawyers Weekly content placements
- Referral from beta firms
- Targeted Google Ads on bottom-funnel terms ("LEAP litigation plugin", "QLD costs narrative software")
- Webinars and CPD-eligible content (lawyers must do 10 CPD points/year — eligible content gets registrations)

---

## 8. Regulatory & Ethical Considerations

### 8.1 Privacy Act 1988 (Cth)

- Plugin handles personal information about clients, opposing parties, witnesses
- APP 1 (open and transparent) — Privacy Policy required
- APP 6 (use/disclosure) — only for the purposes the customer firm authorises
- APP 8 (cross-border) — **avoided entirely** by Sydney-only hosting
- APP 11 (security) — encryption, access control, audit
- APP 12 (access) — customer firm can extract per-matter data on request

### 8.2 Legal Profession Act 2007 (Qld) and Australian Solicitors' Conduct Rules

- ASCR Rule 9 (confidentiality) — solicitor remains primarily responsible; plugin must support, not undermine, confidentiality
- LPA s 8 / Rule 9.2 — disclosure to a third-party processor (LitIQ) must be authorised by client engagement terms; LitIQ provides template engagement clause
- ASCR Rule 4 (integrity to the court) — AI output flagged as draft; solicitor cannot abdicate responsibility
- LPA Pt 3.4 — costs disclosure obligations unaffected; LitIQ assists with bills, doesn't replace disclosure

### 8.3 Client Legal Privilege

- Privilege over communications and documents must not be waived by sending data to a processor
- LitIQ's role: same as outsourced typist or off-site server — agent of the solicitor, not a third-party stranger
- Engagement clause: client consents to confidential disclosure to LitIQ for processing
- DPA: LitIQ contractually bound to confidentiality and to assert privilege if subpoenaed

### 8.4 AI Ethics Posture

- Solicitor remains responsible for all outputs — built into ToS and UI ("Draft only — solicitor review required")
- No AI output exits the plugin without an explicit user export action
- Audit log of AI prompts and outputs for 7 years
- No anthropomorphised "AI lawyer" branding — LitIQ assists, it does not advise

### 8.5 Data Sovereignty

- All data, all inference, all backups in `ap-southeast-2`
- Contractually with Anthropic: no overseas inference, no training on customer data, no retention of prompts/outputs beyond service delivery

### 8.6 Terms of Service

- Plugin provides tools, not legal advice
- No reliance disclaimer for AI output
- Customer indemnifies LitIQ for use of outputs in legal practice
- LitIQ liability cap: 12 months fees paid (typical SaaS)
- Australian governing law (QLD)

### 8.7 Professional Indemnity Exposure Analysis

**For LitIQ:** standalone PI cover required ($5–10M for SaaS) — *not* legal practice PI; technology errors and omissions cover. Estimated annual premium: $8–15k.

**For customer firms:** their existing PI is unchanged. The plugin produces drafts; the solicitor signs off. If anything, structured gap detection *reduces* PI exposure by catching procedural defaults before they cause client loss.

**Open question:** does any QLD Law Society guidance require declaration of AI use in costs disclosure or to clients? Currently no specific rule, but the QLS has issued general guidance on AI in legal practice (March 2024). **Recommendation:** standard engagement clause disclosing AI-assisted processing.

### 8.8 QLS Guidance on AI

- QLS published *Guidance on the use of generative AI in legal practice* (current as of 2024–2025)
- Key principles: solicitor responsibility, confidentiality, supervision, accuracy verification
- LitIQ designed to satisfy each: human-in-the-loop UX, confidentiality contracts, audit trail, citation verification
- Maintain a regulatory monitoring function as part of operations

---

## 9. Intellectual Property

### 9.1 Protectable Assets

| Asset | Protection | Notes |
|--------|-----------|-------|
| Codebase | Copyright (automatic) | Maintain version control, contributor agreements |
| Database schema and content | Copyright + trade secret | Knowledge bank curation is the moat |
| Scoring algorithms (case ranking, evidence) | Trade secret + potential method patent | Patent has uphill in AU but possible for technical methods |
| Prompts and prompt chains | Trade secret | Never disclosed externally; prompt vault encrypted |
| Strategy playbooks | Copyright (literary work) + trade secret | High-value content; license, do not sell |
| Brand name / logo | Trademark | IP Australia application Class 9 + 42 |
| Authority annotations | Copyright (selection/arrangement) | *Telstra v Phone Directories* — protectable selection |

### 9.2 Patent Strategy

- AU and PCT applications worth considering for:
  - **Element-evidence mapping with automated gap inference** — novel pipeline combining cause-of-action elements, evidence scoring, and procedural deadlines
  - **Costs-narrative generation from PMS time entries** — likely too descriptive, but worth a provisional
- Cost: ~$8–15k provisional; $40–80k full PCT; defer until Series A or product-market fit confirmed

### 9.3 Ownership

- Boss Lawyers Pty Ltd retains 100% ownership initially
- Spin-off into a separate `Pty Ltd` entity recommended at MVP launch (see §10)
- Boss Lawyers receives ongoing royalty / equity in NewCo if spun out
- All contractor IP assigned via standard IP assignment deed

---

## 10. Entity Structure Options

### 10.1 Option A — Boss Lawyers Develops and Sells Directly

**Pros:**
- Simplest; no new entity, no new accounting
- Existing R&D tax incentive eligibility (already < $20M turnover)
- Founder retains 100%

**Cons:**
- Mixes legal practice (regulated, PI-insured) with software business (different risk profile)
- Solicitor practice rules complicate equity/options for tech employees
- Hard to sell or raise capital cleanly later — buyer/investor doesn't want a law firm attached
- LPA / ASCR may restrict revenue share with non-solicitors in the same entity
- Dilutes Boss Lawyers' brand if product underperforms

### 10.2 Option B — New Pty Ltd Subsidiary or Sister Co (RECOMMENDED)

**Structure:**
- `LitIQ Pty Ltd` (or `[Brand] Technologies Pty Ltd`)
- Mark holds shares directly or via existing trust structure
- Boss Lawyers Pty Ltd licenses frameworks, knowledge bank seed content, brand goodwill to NewCo via arms-length licence agreement
- NewCo employs/contracts developers, owns SaaS IP, signs customer contracts

**Pros:**
- Clean separation of regulated legal practice from unregulated software business
- Clean P&L for raising capital, granting ESOP, or selling
- Limits PI exposure to NewCo's E&O cover; doesn't bleed into legal practice PI
- ESVCLP / R&D incentives apply cleanly to NewCo
- Brand of NewCo can outgrow Boss Lawyers' brand

**Cons:**
- Setup cost (~$3–8k legal/accounting), ongoing compliance for two entities
- Inter-company agreements need careful drafting (transfer pricing, IP licence)
- Dual filings, dual financials

### 10.3 Option C — Joint Venture with Tech Co-founder

**Pros:**
- Brings technical capacity from Day 1
- Shares execution risk
- Faster build velocity

**Cons:**
- 15–30% dilution to a co-founder Mark may be able to substitute for with contractors
- Co-founder fit risk (highest single point of failure for early-stage)
- Mark loses unilateral product control

### 10.4 Recommendation

**Option B.** Incorporate `LitIQ Pty Ltd` (or eventual chosen brand) as a sister company to Boss Lawyers Pty Ltd, owned by Mark personally or via the existing family trust. Boss Lawyers licenses content to NewCo. Engage a senior contract developer (not co-founder) for MVP. Reserve equity for a future technical co-founder/CTO if needed at Series A, not before.

### 10.5 Tax Considerations

- **R&D Tax Incentive (RDTI)** — 43.5% refundable tax offset for group turnover < $20M; substantial offset on AI engineering and prompt design work; budget $30–60k recoverable annually during build phase. Maintain contemporaneous R&D records.
- **ESVCLP eligibility** — if raising capital, structure raise to allow Early Stage Venture Capital Limited Partnership investors (uncapped CGT exemption for them) to maximise valuation
- **Early Stage Innovation Company (ESIC)** — NewCo likely qualifies; investors get 20% non-refundable carry-forward tax offset and CGT exemption on shares held >1 year. Powerful angel-raise tool.
- **Trust v company:** company structure preferred for SaaS — investor-ready, ESOP-able, R&D-eligible. Family trust as shareholder is standard.
- **Transfer pricing on IP licence** between Boss Lawyers and NewCo — get accountant advice; benchmark licence fees at arms length to satisfy ATO

---

## 11. Development Roadmap

### 11.1 Month-by-Month

| Months | Milestone | Owner | Cost |
|--------|-----------|-------|------|
| M1–2 | LEAP API integration (sandbox), Postgres+pgvector schema, Next.js scaffold, Auth0 integration, base UI | Lead dev + Mark | $15–25k |
| M3–4 | Case Ranking engine, Element Tracker, Knowledge Bank seed (QLD authorities) | Lead dev + Mark | $20–30k |
| M5–6 | Gap Detector, Evidence Inventory, Billing Narrative Generator, MVP launch to 3–5 beta firms | Lead dev + Mark + design contractor | $20–30k |
| M7–8 | Document Drafter v1 (5 doc types), Knowledge Bank UX polish, telemetry | Lead dev | $15–25k |
| M9–10 | Beta program iteration, paid conversion of beta firms, LEAP Marketplace submission | Mark (sales) + dev (bug fixes) | $5–10k |
| M11–12 | LEAP Marketplace approval, public QLD launch, first 20 paying firms | Mark + first CSM (PT) | $10–15k |

**Total MVP build: ~$80–150k outsourced.**
**With Mark contributing significant AI-assisted development: ~$40–80k.**

### 11.2 Key Hires

| Role | When | Type | Cost (annual loaded) |
|------|------|------|----------------------|
| Senior full-stack developer | Month 3 | Contractor → FTE at M9 | $180–220k |
| Designer (UI) | Month 4–5 | Contractor (3 months) | $25–40k |
| Customer Success (PT) | Month 9 | Contractor | $40–60k |
| Customer Success (FT) | Month 14 | FTE | $110–140k |
| Sales (national) | Month 18 | FTE | $150–200k OTE |

### 11.3 Build vs Buy

| Component | Decision | Reasoning |
|-----------|----------|-----------|
| Auth | Buy (Auth0/Clerk) | Commodity |
| LEAP integration | Build | Core IP |
| AI inference | Buy (Anthropic) | Foundational; commoditising |
| Embeddings | Buy (Voyage) | Commodity |
| PMS-side features | Don't build | Stay in lane |
| Knowledge bank curation | Build | Core IP and moat |
| Document generation pipeline | Build | Core IP |

---

## 12. Risk Analysis

| # | Risk | Likelihood | Impact | Mitigation |
|---|------|------------|--------|------------|
| 1 | LEAP changes API or marketplace terms | Medium | High | Build abstraction layer; design for multi-PMS from Year 2 (Smokeball, SILQ, Actionstep); monitor LEAP partner communications; maintain personal relationship with LEAP partner team |
| 2 | LEAP builds competing in-house feature | Medium | High | Move fast; depth of QLD-specific knowledge; ship features LEAP won't (opinionated litigation workflow); become acquisition target if pressured |
| 3 | Big legal-tech entrant (Harvey, Thomson Reuters) targets SME | Medium | High | Beachhead defensibility (LEAP-native, QLD-specific, lawyer-built); price/feature differentiation; incumbents move slow on SME |
| 4 | Regulatory change (AI in legal practice) | Medium | Medium | Active QLS engagement; conservative product positioning ("draft, not advice"); audit trail by design; maintain regulatory monitoring |
| 5 | Low adoption / market too small | Low–Medium | High | Start with highest-pain feature (Billing Narrative Generator) to drive adoption; tier pricing for entry; beta-to-paid gate validates demand before scaling spend |
| 6 | AI hallucination in document drafting | High (without controls) | High | Quality gates: authority verification, cite-check, mandatory human-review checkpoint before export, audit log; clear UI labelling of all AI output as draft |
| 7 | Mark's time split between practice and product | High | Medium | Cap product time at 20% in Year 1; hire dev contractor; productise things that already exist (frameworks); product time = leverage on practice (dogfood); set hard "no after-hours" rule on product |
| 8 | Privacy breach / data leak | Low–Medium | Catastrophic | Sydney-only hosting; encryption everywhere; SOC 2 path; pen-test before launch; cyber insurance ($5–10M); incident response plan |
| 9 | AI provider (Anthropic) outage or price hike | Medium | Medium | Multi-model abstraction layer (Claude primary, OpenAI / open-source fallback); cost controls per-tenant; price tiers reflect cost reality |
| 10 | Founder dependency (bus factor = 1) | Medium | High | Document everything (frameworks, prompts, playbooks); hire dev who can own product on Mark's behalf; aim for 2-key processes by Year 2 |
| 11 | Customer churn (low stickiness) | Medium | Medium | Annual contracts (with discount); deep LEAP data integration creates switching cost; firm-specific knowledge bank additions create lock-in |
| 12 | Slow LEAP Marketplace approval / rejection | Medium | Medium | Engage LEAP partner team early (Month 1); design to LEAP partner standards from Day 1; have direct-to-firm fallback if marketplace path stalls |
| 13 | PI / professional negligence claim arising from AI output | Low | High | E&O insurance; ToS liability cap; mandatory human-review UX; audit log proves draft-only nature |
| 14 | Rapid commoditisation of AI legal tools | High | Medium | Compete on workflow depth and AU-specific knowledge, not raw AI; lock in customers via integrations and firm-specific data |
| 15 | Talent risk (can't hire experienced AU dev) | Medium | Medium | Use AU contractor market; consider EU/UK contract developers (timezone-friendly); upskill via AI-assisted dev; pay market rate |

### 12.1 Top-3 Existential Risks (in order)

1. **LEAP relationship** — single point of failure. Mitigation: multi-PMS by Year 2.
2. **Founder bandwidth** — Mark cannot run a litigation practice and a SaaS company simultaneously without dedicated build help. Mitigation: hire dev contractor by Month 3, no exceptions.
3. **Competitive entry by deep-pocketed incumbent** — Mitigation: speed to market, depth in QLD/AU, brand.

---

## 13. Decision Points / Open Questions

For Mark to resolve before initiating build:

- [ ] Confirm name and lodge IP Australia trademark application
- [ ] Decide entity structure (Option B recommended) and engage accountant
- [ ] Validate LEAP Marketplace partner program terms (partner agreement, revenue share, technical requirements)
- [ ] Confirm capital strategy (bootstrap vs raise) — recommendation: bootstrap to $215k ARR, then evaluate
- [ ] Recruit 3–5 beta firms (informal commitments before build starts)
- [ ] Engage senior full-stack contractor for MVP (Month 1–2 hire)
- [ ] Engage cyber/PI insurance broker for E&O quote
- [ ] Engage QLS or external counsel on AI use disclosure in client engagement terms
- [ ] Build Pricing/ROI calculator for sales conversations
- [ ] Confirm AI inference region availability with Anthropic (Sydney endpoint or contracted no-retention routing)

---

## 14. Appendices

### Appendix A — Referenced Internal Frameworks

- `reference/case-ranking-framework.md` — Merit/Risk/Commercial scoring rubric
- `reference/evidence-assessment-framework.md` — 5-dimension evidence scoring
- `reference/knowledge-bank-architecture.md` — Semantic search and authority retrieval design
- `reference/document-generation-spec.md` — Drafting pipeline and quality gates
- `reference/qld-pleading-guide.md` — UCPR pleading standards (informs Document Drafter)
- `reference/qld-pleading-authorities.md` — Canonical authorities (knowledge bank seed)
- `TOOLS.md` — QLD costs narrative principles (Billing Narrative Generator spec)

### Appendix B — Glossary

- **LEAP** — Legal practice management software dominant in AU SME market
- **MVP** — Minimum viable product (v1.0)
- **ARR** — Annual recurring revenue
- **CMEK** — Customer-managed encryption keys
- **DPA** — Data Processing Agreement
- **ESVCLP** — Early Stage Venture Capital Limited Partnership
- **ESIC** — Early Stage Innovation Company
- **RDTI** — R&D Tax Incentive
- **UCPR** — Uniform Civil Procedure Rules 1999 (Qld)
- **APP** — Australian Privacy Principle
- **ASCR** — Australian Solicitors' Conduct Rules
- **LPA** — Legal Profession Act 2007 (Qld)
- **PI** — Professional Indemnity (insurance)
- **E&O** — Errors & Omissions (insurance)

### Appendix C — Document Control

- Author: Mark Harley
- Status: Draft v1.0 — for internal review
- Distribution: Confidential. Do not distribute without author's consent.
- Next review: After beta firm recruitment confirms market signal.

---

*End of brief.*