Case Search & Discovery: Enterprise Investigation Retrieval

Overview#

An analyst working a vehicle theft case notices a name that rings a bell. She suspects it appeared in a separate handling investigation two years ago, but she cannot recall the case number. In a traditional system, verifying that hunch means emailing colleagues, waiting for replies, and hoping someone remembers. In the platform, she types the name and finds the connection in seconds, along with every other case where that entity appears.

The Case Search & Discovery module transforms investigation retrieval from fragmented manual lookups into fast, comprehensive full-text searches across large case repositories through advanced indexing, filtering, saved search templates, and relationship mapping. It enables investigators to instantly find relevant cases, discover hidden connections, and draw on institutional knowledge through search capabilities specifically designed for complex investigations requiring pattern recognition, precedent analysis, and intelligence sharing.

Key Features#

• Natural Language Search: Investigators type questions or descriptions as they would speak, and the system interprets intent to deliver relevant results without requiring knowledge of query syntax
• Advanced Filtering: Refine search results by date ranges, case types, investigators, jurisdictions, evidence types, custom metadata fields, and other attributes for precise discovery
• Saved Search Templates: Preserve complex queries as one-click templates for recurring search needs, improving efficiency and ensuring consistency for common lookup patterns
• Smart Folders: Automatically populate folders with cases matching dynamic criteria, keeping curated collections current without manual updates as new cases are created
• Relationship Discovery: Map connections between cases based on shared entities, addresses, contact information, financial transactions, and behavioural patterns to reveal hidden links aligned with the POLE model
• Boolean and Field-Specific Queries: Support precise search construction with Boolean operators and targeted field searches for advanced users who need granular control
• Fuzzy Matching: Account for typos, name variations, transliterations, and spelling differences when searching across case records to prevent missed results
• Real-Time Indexing: New cases, notes, and evidence appear in search results within seconds of creation, ensuring investigators always search the most current data
• Search Result Ranking: Results are ranked by relevance using multiple factors including match quality, recency, case importance, and personalisation based on the searcher's role and history
• Cross-Case Pattern Analysis: Identify recurring patterns, common methodologies, and correlated activities across case populations through aggregate search and analysis capabilities

Use Cases#

Precedent Analysis#

Investigators search for similar past cases to identify proven investigation strategies, common patterns, and lessons learned that can inform current investigations. Historical case outcomes provide guidance for approach selection and resource allocation.

Pattern Recognition#

Analysts use relationship mapping and cross-case searches to identify connections between seemingly unrelated cases, uncovering networks, repeat offenders, and coordinated activity that would be invisible when examining cases individually.

Rapid Case Lookup#

Front-line staff quickly retrieve specific cases using natural language, reference numbers, or key identifiers, reducing time spent navigating case hierarchies and file structures to find the right record.

Intelligence Sharing#

Teams share saved searches and smart folders to collectively monitor case populations matching specific criteria, enabling proactive investigation management and early identification of emerging trends.

Deconfliction#

Before initiating new investigations, teams search for existing cases involving the same subjects, locations, or activities to prevent duplicate investigations and identify opportunities for coordination.

Workflows#

Investigative Search#

1. Enter a natural language query, keywords, or structured search criteria in the search interface
2. Review results ranked by relevance with highlighted matches and contextual previews
3. Apply filters to narrow results by date, type, jurisdiction, status, or other attributes
4. Open cases directly from search results to review details and take action
5. Save effective searches as templates for reuse or share with team members

Relationship Exploration#

1. Start from a case, entity, or data point of interest
2. The system identifies related cases and entities based on shared attributes and connections
3. Explore the relationship map to discover linked cases, shared entities, and connecting patterns
4. Drill down into specific connections for detailed context and supporting evidence
5. Save the relationship view for ongoing monitoring or export for further analysis

Smart Folder Management#

1. Define the criteria for a smart folder based on case attributes, date ranges, and other filters
2. The system automatically populates the folder with all matching cases
3. New cases matching the criteria are added automatically as they are created
4. Review the smart folder regularly to monitor the case population and identify changes

Integration#

• Programmable API Access: Execute searches and retrieve results programmatically for integration with external analysis tools, investigation platforms, and custom applications
• Business Intelligence Tools: Feed search results and relationship data into BI platforms for advanced visualisation, cross-system correlation, and trend reporting
• Case Management Systems: Embed search capabilities within broader case management workflows for seamless access to discovery features
• Alert Services: Configure saved searches to generate automatic alerts when new cases matching specified criteria are created or when relationship patterns change

Open Standards#

• GraphQL (June 2018 Specification): All search queries, saved-search mutations, and faceted result retrieval are exposed through a GraphQL API built with the Strawberry library, conforming to the GraphQL specification.
• POLE Data Model (UK National Intelligence Model): Relationship discovery and cross-case entity mapping are structured around the Persons, Objects, Locations, and Events (POLE) framework used in law enforcement intelligence practice.
• BM25 / ISO SQL Full-Text Search (PostgreSQL tsvector): Full-text indexing uses PostgreSQL to_tsvector and ts_rank, implementing BM25-style relevance scoring for keyword and phrase retrieval across case records.
• Reciprocal Rank Fusion (RRF, Cormack et al. SIGIR 2009): Hybrid search combines BM25 keyword results and dense semantic vector results using the RRF algorithm (k=60 smoothing constant) to produce a single fused relevance ranking.
• ISO 8601: Date-range filter parameters passed to the search engine are serialised and parsed as ISO 8601 date-time strings throughout the search and indexing layer.
• RFC 4122 (UUID): Every case record, search index, saved-search template, and indexing job is identified by a version-4 UUID, ensuring globally unique, collision-resistant identifiers.
• JSON Web Token (RFC 7519): All search API operations require a valid RS256-signed JWT issued by the platform's authorisation service; access is refused without a verified bearer token.

Last Reviewed: 2026-02-05 Last Updated: 2026-04-14