Insurance & Compliance
AI-Powered Legal and Regulatory Summarisation Platform
Major Insurer
Timeline: 7 months
Team: 5-7 specialists
KEY IMPACT
Plan for reduced compliance review time, increased accuracy and regulatory adherence, and established automated knowledge reuse across departments.
The Challenge
A major insurer was drowning in regulatory text. Between APRA prudential standards, ASIC consultation papers, AUSTRAC AML/CTF guidance, state-based insurance acts, and a constant stream of industry circulars, the legal and compliance team was reviewing hundreds of pages per week and trying to translate them into actionable internal guidance for product, claims, and underwriting teams.
The existing process was almost entirely manual: senior compliance analysts read each new regulation, drafted internal summaries, distributed them via email, and fielded follow-up questions. Knowledge silos formed quickly — the underwriting team's interpretation of a circular often differed from claims's, and there was no single canonical summary that could be relied on for audit purposes. Audit cycles repeatedly surfaced delays and inconsistencies that were starting to attract regulator attention.
The insurer had also grown organically through several acquisitions, so the legal corpus included historical interpretations from multiple legacy entities that were not always consistent with current group policy. Compliance staff routinely had to dig through archived files to figure out which interpretation was authoritative, which made even simple lookups slow and error-prone.
Leadership wanted a solution that could dramatically accelerate the summarisation process without ever creating the kind of unverified output that would itself become a compliance risk. The bar was high: any AI-generated summary had to be demonstrably faithful to the source, version-controlled, and reproducible on demand for audit.
Our Solution
We helped plan and build a Databricks-based summarisation workflow that combined RAG, transformer-based text generation, and ROUGE-based evaluation pipelines into a single governed platform.
Documents were vectorised on ingestion and stored in a scalable Delta Lake repository with full version history. Every regulatory text entered the system with metadata covering its source authority, effective date, supersession status, and applicability to specific business lines, enabling fine-grained retrieval and authoritative-source resolution.
Using LangGraph orchestration, the system dynamically adjusted summarisation depth depending on document type and compliance category. Short circulars used an extractive style that produced near-verbatim digests with citations. Longer prudential standards used an abstractive style that synthesised the document into a structured brief — purpose, key obligations, affected business lines, recommended actions, and open interpretive questions — that compliance staff could review and finalise in a fraction of the time.
Model outputs were continuously evaluated via DeepEval and tracked using MLflow Evaluate, creating an evidence-backed audit trail for every generated summary. Each summary carried a faithfulness score, links to the underlying document chunks, and the model version that produced it. The compliance team could re-run any summary on demand and compare outputs across model versions — a capability that proved decisive during their next audit cycle.
Integration with existing compliance workflows ensured every model output could be versioned, reproduced, and revalidated. New summaries flowed automatically into the compliance team's existing review queue, with high-confidence items pre-approved and low-confidence items routed for SME attention. The platform did not replace human judgment; it removed the manual drudgery so that experts could spend their time on the genuinely ambiguous interpretive questions where they add the most value.
AI-Powered Legal and Regulatory Summarisation Platform Architecture showing RAG orchestration, document vectorisation, summarisation engine, and cross-department knowledge library
Results & Outcomes
Plan for reduced compliance review time across regulatory updates and circulars
Increased accuracy and regulatory adherence through evaluated, traceable summaries
Established automated knowledge reuse across underwriting, claims, and product departments
Provided audit-ready evidence trail for every AI-generated compliance artefact
Technologies Used
Databricks
LangGraph
Vector Search
MLflow Evaluate
DeepEval
Delta Lake
Ready for Similar Results?
Let's discuss how we can help transform your organisation's data and AI capabilities.