Was any client data involved?

No. This run answers a general point of law with no client or personal data. The PII Redaction step found nothing to redact, a good confidentiality check, so the demo is safe to share as-is.

What is the Citator, and why does it matter?

The Citator runs an independent good-law check on every case and assigns a treatment signal. It's why the memo honestly marks the recent Sony authority 'unverified' rather than overclaiming, and why two governance checks are flagged for a human to confirm before you rely on the result.

What do I get by email?

The full memo as a PDF, sent to the address you enter before the demo. We use your details only to send the memo and follow up about AI governance, no spam, and you can opt out anytime.

Watch a Research Agent run: interactive demo

Interactive demo

Watch a real governed research run

This is a replay of an actual Research Agent run, not a mock. Approve the plan, watch the specialist agents and governance gates work, then read and download the memo the agent produced.

Research Agent

New research task

Are damages-based litigation funding agreements enforceable after PACCAR?

GPT-5.4Prepared for the client's board · WordEngland & Wales

Run this query

Add your details to watch a real governed run, gate by gate, then download the full memo. We'll email you a copy too.

What you just watched

One orchestrator, a fleet of specialist agents, all governed

Every stage in the demo maps to a real part of the platform: named sub-agents doing the work, wrapped by the governance layer that verifies it and flags what it can't.

Preparation

Pre-run guardrails that run before a single source is touched.

Planning Protocol
Builds a structured execution plan for the legal task.
PII Redaction
Strips client PII from prompts before they leave the tenancy.
Jurisdiction Gatekeeper
Blocks cross-jurisdictional contamination, such as Scottish authority on an England and Wales matter.

Research

A fleet of specialist sub-agents, each named and visible.

Query Planner
6 dependency-tagged sub-questions + 3 clarifications
Source Identifier
Mapped each plan item to open-web sources
Primary-Law Retrieval
On-point passages with verbatim quotes and pinpoints
CitatorFlagged
PACCAR explained; Sony treatment unverified
Synthesis & Drafting
Research memo + board-level Word briefing produced

Checks

Post-run governance that proves, or flags, the result.

Privilege Review
Scans output for inadvertent privilege or confidentiality leaks.
Audit Trail
Generates the immutable compliance certificate.
Source Freshness Check
Flags overruled, repealed or superseded authorities.
Citation VerificationFlagged
Independent re-fetch and verbatim quote match. The Research anti-hallucination gate.
Research Confidence ScoringFlagged
Per-section low, medium or high with deterministic floor rules.
Legal Citation Verifier
Verifies citations against BAILII, legislation.gov.uk and EUR-Lex.
Citation Style Formatter
OSCOLA, OSCOLA-Ireland, Bluebook or house style.

Every run is wrapped by 13 governance protocols and 6 deterministic gates that turn a draft into something a firm can rely on.

See how governance works

See it run on your own question.

Book a 30-minute pilot call with founder Daman Kaur. One-week guided onboarding, your data stays in the UK or EU, and every run is governed from day one.

Book a pilot call Join the waitlist

Questions

The demo, answered.

How verification, governance, and the agents work inside an SRA-regulated firm.

Still have questions? Talk to the founder

Yes. It's a replay of an actual Research Agent run on the enforceability of damages-based litigation funding after the Supreme Court's PACCAR decision. The plan, the citations, the governance flags, and the memo are exactly what the agent produced.