The problem
The firm managed monthly debtor lists from a banking client on an unstable, locally hosted system. Downtime and inconsistent data were routine, and every letter meant copying fields into a Word template by hand, for four different loan products with different rules.
- Debtor lists arrived monthly as Excel and were processed manually
- The legacy system suffered frequent downtime and data inconsistencies
- Placeholders in letters were filled inconsistently from one clerk to the next
- Mortgage, credit card, hire purchase and personal loan cases each needed different documents and metadata
What we built
A cloud platform that covers the full recovery lifecycle. Civil process: upload the bank's workbook, get typed debtor records, generate notices, summonses, affidavits and bankruptcy papers from templates the firm maintains itself. Execution proceedings: writ of seizure and sale and prohibitory order ledgers, with post-judgment interest computed exactly and exported in the bank's own statement format, for both conventional and Islamic facilities.
- Encrypted Excel workbooks are parsed into validated records, with a per-row error report
- Letters render from the firm's Word templates; any variable that could not be filled is flagged on the document record
- An interest engine produces statements of account to the cent, as Excel and PDF
- AI reads statements of account into structured figures for a person to confirm
- Cases are assigned to teams, with progress, audit history and a document register
- Invoices are created in SQL Accounting directly from the litigation record
Built as a platform
Clausework is multi-tenant from the data layer up. Every record belongs to an organisation, every query is scoped to it, and guard tests fail the build if a new model or endpoint forgets. Roles and access rights are configurable per firm, and support staff can view a tenant in read-only mode without being able to change anything.
Outcome
The firm retired the legacy system. Filing and checking time fell by 95%, documents go out complete, and managers can see where every case stands.
- No more dependence on a single office machine
- Structured handling of every debtor and letter type
- A full audit trail for the bank and for the firm's own oversight
- React 19
- TypeScript
- Mantine
- Django 6
- PostgreSQL
- docxtpl
- Google Gemini
- AWS App Runner
- S3
- Terraform