LegalDocs-Assist — Data & AI Summary

LegalDocs-Assist is a cloud‑hosted, AI‑assisted drafting platform for UK solicitors, paralegals and law firm partners working in RTA (including credit hire), Employers’ & Public Liability, housing disrepair, and clinical negligence. It helps fee-earners analyse correspondence, draft replies to opposing correspondence, generate CPR-compliant witness statements, and work from a curated case law base — with server-side safeguards designed to keep client data out of model training and unverified citations out of finished drafts.

How We Handle Your Data

• Browser‑side processing of source documents: Text‑based PDFs and Word documents are processed in your browser; the original file does not leave your device and only the extracted text is sent to our server. Two narrow exceptions: scanned/image‑only PDFs are briefly uploaded for OCR and deleted within the hour; medical pagination uploads the PDF (see Retention below).
• Server‑side PII masking gateway: Before any text reaches the AI model, it passes through a mandatory server‑side masking layer covering 14 verified categories: personal names (titled names always; bare / untitled names in high-precision contexts such as field labels, salutations and sign-offs), email addresses, UK phone numbers, National Insurance numbers, NHS numbers (Mod 11 validated), UK driving licence numbers, vehicle registration marks (VRMs), IBANs, credit/debit card numbers (Luhn-checked), sort codes, bank account numbers, dates of birth (in context), policy and claim references, and street addresses. Detected items are replaced with neutral placeholders. An additional aggressive output-side re-scrub catches bare First-Last name pairs in AI responses, with guards for case citations and known organisation names. Masking is pattern‑based and not a perfect filter — users are asked to send only what the task needs. UK postcodes are deliberately preserved (material to BHR locality arguments and venue analysis), protected by enterprise no‑training terms and an output‑side scrubber that strips any postcode the AI introduces that was not in the input.
• AI runs on our server, output deleted, not retained: The AI drafting and generation happens on our server (that is where the model runs) under Zero Data Retention terms with the AI provider. The generated output is produced on our server and then deleted, not retained.
• Names in the Witness Statement Generator: In the Witness Statement Generator, the party names you enter on the form (claimant, defendant, witnesses) are passed to the model so the statement attributes evidence to the correct person. In the analysis tools, titled and labelled names are masked like any other identifier.
• Zero Data Retention at the AI provider: The AI provider operates under the Lovable AI Gateway's Zero Data Retention terms, in force since March 2026. Your prompts and the model's responses are not retained by the provider and are not used to train any model.
• Encrypted, tenant‑isolated storage: Your data is stored in an encrypted database with row‑level security. Each company operates in its own completely separate data partition — there is no cross‑company access.

Retention & Deletion

• Retention: LegalDocs Assist is designed not to retain personal case data. Uploaded source documents and the outputs the tool generates (witness statements, paginated bundles, chronologies, timelines, advice, correspondence) are deleted shortly after they are produced or on session close, and a scheduled purge runs continuously as a backstop so case content is never retained beyond 24 hours. Firms download or save their own copy and re‑run if needed. Medical‑bundle pagination files are deleted within roughly one hour of completion. Inactive login sessions expire automatically. The only data retained is the minimal account record needed to run the subscription; activity and audit logs are retained as described in the Privacy Policy.
• Backups: Uploaded files are never included in any backup. The database has standard disaster‑recovery backups, but because case content is deleted within 24 hours those backups hold essentially no personal case data, and once a record is deleted there is nothing to restore it from.
• Special category (health) data: The Service routinely processes special‑category health data (medical records, injury, treatment) for clinical‑negligence and personal‑injury work. Safeguards: browser‑side processing of source documents, server‑side PII masking, encryption, and immediate deletion within 24 hours. As data controller, your firm is responsible for the Article 9 lawful basis.
• Deletion: Deletion is immediate and permanent. When a record or document is deleted it is removed from the live database and, where it has associated files, from storage.
• Firm-level erasure: On contract termination, or on a verified erasure request, we permanently delete your firm's tenant — all case content and associated storage objects, together with the user accounts. The action is executed only by CaseFlow Automation Ltd platform administrators via a hardened admin path (archive first as a reversible step for thirty (30) days as a grace period, then permanent delete from the archived list as a second gate); firms cannot self-delete. Where erasure is triggered by contract termination it is completed within thirty (30) days. It is recorded in a deletion audit log that holds only the fact of deletion, who performed it, when, and the counts — never the deleted content itself. This is a contractual commitment honoured operationally, not an automated platform feature.
• Data subject rights: Under UK GDPR you have rights of access, rectification, erasure, restriction, portability and objection. For claimant data your firm is controller and we are processor; we support requests on your instruction. For account data we hold as controller, contact info@caseflowautomation.co.uk.

How We Use AI

• Decision‑support only: AI analyses correspondence and drafts responses, but humans always review and approve before anything is sent. Every output is labelled as a draft requiring review.
• 3‑layer anti‑hallucination guard: Every AI output passes through three independent safety layers: (1) strict prompt rules prohibiting the AI from citing cases outside our knowledge base of 210 UK authorities, (2) an automated citation validator that cross‑references every case name against our case law database and strips citations that are not in it, and (3) a text‑body scrubber that scans all prose for remaining case‑name patterns and removes any that cannot be verified.
• No AI training on your data: We do not use your data to train AI models. Our AI provider operates under enterprise terms that prohibit training on your inputs and outputs.

Key Commitments

Full Documentation

This page is a summary. Full details are set out in:

• Privacy Policy — our formal data processing commitments under UK GDPR
• AI Safety & Security Policy — governance, anti‑hallucination controls, and model selection
• How We Protect Your Data — our two‑layer privacy architecture explained
• How It Works — Plain English Guide — a jargon‑free overview for non-technical users