Preparing for a hearing — whether a directions appointment, a trial, an employment tribunal or a judicial review — is one of the most time-intensive tasks in litigation practice. AI can dramatically reduce that time. But doing it right requires understanding both the opportunity and the risks.
The opportunity: An AI that has read your entire case file can produce a structured pre-hearing brief — facts summary, key issues, leading authorities, weaknesses, checklist — in seconds. A task that takes a junior associate two hours can be done before your morning coffee.
1. What a good AI pre-hearing brief contains
A well-structured AI-generated hearing brief should cover:
- Case summary — parties, background facts, procedural history, hearing type and date
- Key legal issues — the precise questions the court or tribunal must decide
- Our position — client's case on each issue, with references to supporting documents
- Leading authorities — the most relevant cases and statutory provisions
- Adverse authorities — cases the other side will rely on and how to distinguish them
- Weaknesses and risks — candid assessment of the vulnerabilities in the client's case
- Missing evidence or disclosure — flagging gaps in the bundle that could be exploited
- Hearing checklist — documents to bring, procedural steps, time estimates
- Confidentiality marking — the brief must be clearly marked as privileged and not for disclosure
Standard structure — LegisBox pre-hearing brief
PRE-HEARING BRIEF — [CASE NAME] — [COURT/TRIBUNAL]PRIVILEGED AND CONFIDENTIAL — NOT FOR DISCLOSURE
─────────────────────────────────────────
Hearing: [Date] · [Court] · [Judge if known]
Parties: [Claimant] v [Defendant]
─────────────────────────────────────────
1. KEY LEGAL ISSUES
2. OUR POSITION
3. LEADING AUTHORITIES
4. ADVERSE AUTHORITIES — DISTINGUISHED
5. RISKS AND WEAKNESSES
6. MISSING DISCLOSURE / GAPS
7. PRE-HEARING CHECKLIST ☐
⚠️ CRITICAL POINTS
─────────────────────────────────────────
Internal document · Not for disclosure
2. The LPP issue with cloud AI briefing tools
Pre-hearing briefing notes are almost always privileged — they are prepared in anticipation of litigation and contain the solicitor's legal analysis. This makes them among the most sensitive documents in a case file.
Uploading a litigation brief or the underlying case documents to a cloud AI tool — ChatGPT, Copilot, Harvey AI — involves disclosing privileged material to a third party. As noted in our guide on LPP and ChatGPT, this disclosure can waive LPP in relation to those documents, permanently and irreversibly.
Opposing counsel who discovers that privileged documents were uploaded to a third-party AI service could apply for disclosure of those documents on the basis that privilege has been waived. Courts have taken a strict approach to implied waiver of LPP.
⚠️ The waiver risk is not theoretical: In R (Jet2.com Ltd) v Civil Aviation Authority [2020] EWCA Civ 35, the Court of Appeal confirmed that accidental disclosure of privileged documents to a third party can waive privilege if not remedied promptly. Uploading to an AI is deliberate, not accidental — the risk of waiver is correspondingly higher.
3. What AI can and cannot do for hearing preparation
What AI can do well
- Summarise a large case file into a concise factual chronology
- Identify key documents and cross-reference them across witness statements and correspondence
- Search official legal databases (BAILII, legislation.gov.uk) for relevant authorities
- Draft a structured brief in the format you specify
- Flag inconsistencies in the evidence or gaps in disclosure
- Generate a pre-hearing checklist based on the procedural stage and court rules
What AI cannot do
- Exercise the professional judgement of an experienced litigator — all AI outputs must be reviewed
- Predict how a specific judge will approach a case (though it can analyse their published judgments)
- Guarantee the currency of case law — verify all citations before relying on them
- Replace the tactical assessment of which arguments to lead and which to hold back
4. CPR compliance for AI-assisted preparation
The Civil Procedure Rules impose obligations on solicitors that AI use does not displace:
- CPR 1.3 — parties must help the court further the overriding objective. AI-generated submissions that are verbose, repetitive or poorly targeted undermine this obligation.
- CPR 22 — statements of truth must be signed by the party or their legal representative. An AI cannot sign a statement of truth, and the solicitor signing must have personally verified its accuracy.
- Practice Direction 57AC (witness statements) — prohibits witness evidence that is drafted or heavily influenced by lawyers. AI should not be used to draft witness statements, only to analyse existing ones.
5. The self-hosted approach
The compliance-first approach to AI hearing preparation uses a self-hosted AI installed within your firm's network. This means:
- All case file documents are analysed locally — no disclosure to any third party
- LPP is maintained absolutely — no third party ever touches privileged material
- Legal research queries official databases (BAILII, legislation.gov.uk) rather than relying on the AI's training data
- The brief is generated and stored within your firm's controlled environment
Pre-hearing briefs in seconds — LPP intact
LegisBox reads your case file, searches BAILII and legislation.gov.uk, and generates a structured hearing brief — all within your firm's network. No data leaves. Privilege maintained.
Book a demo →