Ask across the matter
Put a bounded question to the governed matter and review the source-linked material behind the answer.
Source-linked AI for substantial documentary matters
AI legal investigation you can inspect.
LegalRAG Pro uses AI across the governed matter to analyse documents, reconstruct chronology, test propositions, surface supporting and contradictory evidence, expose gaps and propose bounded next work — while keeping the source route and professional decision visible.
Analyse the recordTest propositionsFind contradictionsExpose gapsPropose next work
Synthetic demonstration material. No client information is shown.
31 sec | quickest introduction
AI inside the matter
The AI is part of the working system, not a separate chatbot.
LegalRAG Pro applies AI within the governed matter context. The same documentary record, issues, chronology, people and task state remain available around the AI-assisted work, so the practitioner can investigate rather than simply accept generated text.
Analyse a proposition
Test what supports a proposed account, what qualifies it, what contradicts it and where the record remains uncertain.
Reconstruct chronology
Use AI assistance to organise events and sequence while keeping disputed, approximate or conflicting dates visible.
Run governed investigations
Use Case Operator for broader, multi-step work from a defined objective or approved matter task, with scope and acceptance kept explicit.
Surface gaps and next work
Turn missing information and unresolved questions into visible investigative work instead of filling gaps with invented certainty.
Carry work into drafts and reports
Move reviewed AI-assisted analysis into working drafts and frozen report views while keeping generated work separate from professional approval.
AI power with professional authority
AI accelerates the investigation. The lawyer controls the case.
The AI can
- analyse and compare source material
- surface supporting, qualifying and contradictory evidence
- organise chronology and people around a question
- identify evidential gaps and unresolved questions
- propose bounded next work
- generate working analysis for review
The lawyer decides
- what question and scope matter
- what evidence is complete and legally significant
- whether an inference is justified
- which task or investigation is approved
- what wording is adopted
- what can be relied on, circulated or filed
Source linkage improves inspectability; it does not make an AI conclusion automatically correct.
SRA regulatory context
AI in legal work carries professional obligations.
The SRA's 17 August 2026 warning notice puts particular emphasis on inaccurate AI output, client confidentiality, human oversight, governance and continuing professional responsibility. LegalRAG Pro is designed so those concerns are visible in the working method rather than added as an afterthought.
Risks the SRA highlights
What a legal team has to control.
- False or incorrect AI output, including legal citations.
- Confidential client information and privilege.
- Appropriate supervision, oversight and governance.
- Responsibility for the work remains with the solicitor and firm.
LegalRAG Pro control model
How the workflow responds.
- Source-linked evidence and routes back to the underlying material.
- Bounded AI tasks with uncertainty and limitations visible.
- AI-assisted work kept separate until an explicit professional decision is made.
- Confidentiality, provider and retention controls considered as part of deployment.
Responsible legal AI Read the SRA warning
LegalRAG Pro does not claim SRA approval, certification or automatic compliance.
Finance inside the matter
Financial evidence should be as inspectable as the legal evidence.
LegalRAG Pro Finance brings historical financials, comparable-company analysis, deterministic calculations, source evidence, limitations and report export into the governed matter. The language model is not used as the calculator.
Deterministic calculations
Revenue growth, EBITDA margin, equity value, net debt, enterprise value and multiples are calculated in code from governed inputs, with formula and input trace retained.
Comparable analysis with status
Review the comparable matrix, peer statistics and target position while preserving source conflict, stale-data and assumption-required states.
Constrained Finance AI
Ask against frozen Finance authority without allowing the model to recalculate values, invent missing metrics or turn uncertainty into false precision.
Confidentiality & data protection
Protect the matter before AI work begins.
LegalRAG Pro uses layered controls for confidential documentary work and states the remaining assurance limits openly.
Encrypted storage & backup
Production matter storage is encrypted at rest. Encrypted off-site backup is in place, and an isolated restore of backed-up application data has been successfully verified.
Controlled access
TLS protects traffic to the application. Authentication, an application admission allowlist, matter membership and roles are used to restrict who can reach each matter.
AI provider boundary
LegalRAG Pro sends governed OpenAI API requests with store=False, and API training use is disabled. Under the current standard profile, abuse-monitoring logs may retain customer content for up to 30 days. Zero Data Retention has been requested from OpenAI but is not yet approved or claimed.
Confidential or privileged client use remains subject to a controlled pilot, appropriate customer data-protection arrangements and independent security assurance. No Zero Data Retention claim is made until OpenAI approval is enabled and evidenced.
Choose how much time you have
See the same promise at different depths.
Choose the shortest route to see source-linked work, professional review and the connected matter model.
31 seconds
Quick orientation
Understand the proposition before you read anything else.
Watch the teaser →
90 seconds | recommended
Core workflow
The best first view for partners, litigators and innovation leaders.
Watch the overview →
3 minutes 36 seconds
Architecture & governance
Evidence, professional control and governed Document Lineage.
Watch the architecture film →
8 minutes 58 seconds
Full connected-matter story
See the broader workflow from documentary record to professional review.
Watch the flagship film →90-second case challenge
You inherit 40 documents and four live issues. Where do you look first?
Choose an issue. LegalRAG does not reward the “right” answer; it shows what the current evidence state says and what still needs professional attention.
Evidence incomplete
The missing dates become work, not invented certainty.
The synthetic matter records missing ACAS and claim-presentation dates as an explicit task. The issue is visible as incomplete rather than being smoothed into a confident answer.
Disputed
The chronology is useful because it preserves disagreement.
The record moves through occupational-health advice, adjustment discussions, attendance requirements, capability concerns and disputed meeting material. The task is to inspect what supports, qualifies or contradicts the proposed account.
Comparatively well supported
Support is not the same as automatic legal correctness.
The frozen synthetic report records stronger support for employer knowledge than for some other issues. LegalRAG keeps the evidence route available so the solicitor can verify what the assessment rests on.
Synthetic demonstration status only. These are working evidential positions, not legal determinations.
Ask better case questions
Use AI for investigation, not just drafting.
What actually happened — and when?
Build and inspect a chronology that keeps the evidential route visible.
Where do the accounts stop agreeing?
Keep qualifying and contradictory material visible instead of flattening the record into one narrative.
What should be here but isn’t?
Turn missing information into explicit work rather than filling the gap with model inference.
Show me why you think that.
Move from a proposition back to inspectable source material.
What should I investigate next?
Use bounded tasks and Case Operator work rather than open-ended prompt wandering.
What am I prepared to rely on?
Keep AI-assisted work distinct from professional approval and the case position.
From case-file sprawl to an inspectable case picture
The difference is not more AI text. It is a clearer route through the evidence.
Without a connected matter view
- PDFs, emails, notes and chronology spreadsheets
- Competing versions of the same event
- Important gaps buried in working memory
- AI answers that can look more settled than the record
With LegalRAG Pro
- What happened?
- What supports it?
- What contradicts or qualifies it?
- What is missing?
- What should I inspect next?
Professional control
AI work stays separate until you decide what belongs in the case.
LegalRAG Pro is designed to distinguish AI-assisted findings from the approved professional assessment. A source reference helps you inspect the work; professional judgement determines whether to rely on it.
Explore responsible legal AIWhat should I investigate next?
Turn an evidential gap into bounded work.
The Case Operator is designed around an approved matter task or a defined investigation objective.
Examine the scope, the returned work and the professional decision that follows. Test the complete workflow in a practitioner walkthrough.
Inspect the Case Operator evidence →Start with a concrete evidence question.
For example: identify material that qualifies or contradicts a proposed chronology. Keep the task boundary and the limitations of the available record visible.
Illustrative task, not a representation of a completed product run.
Evaluate the workflow and the safeguards.
How the work connects
Follow one matter from documents through investigation and review.
Trust & due diligence
Review the questions to settle before confidential matter data is introduced.
A defined practitioner pilot
Agree a use case, information boundaries and evidence for an adoption decision.
Try it without client material
Investigate the synthetic matter for yourself.
Start with Eleanor Morgan: a fictional 40-document matter built to expose chronology, disputed issues, gaps and source verification. Then decide whether a scoped practitioner pilot fits your team.