Trovix AI Integration

Trovix Aria ©

AI knowledge retrieval integrated into your document management system

Aria is a retrieval-augmented generation (RAG) system that connects to your existing document management infrastructure — iManage, NetDocuments, SharePoint or any system with an API — and enables fee-earners, underwriters or analysts to query your firm's own knowledge base in natural language. Every answer is grounded in your own documents and includes a citation to the specific source. Aria does not generate from general knowledge. If the answer is not in your indexed content, it says so.

Systems Aria integrates with

Aria connects via API or native connector to your existing DMS and knowledge infrastructure. No data migration. No new system for staff to learn.

iManage Work 10
Native connector. Aria indexes your entire matter library. Search across millions of documents with sub-second response times.
NetDocuments
REST API integration. Full matter and client file indexing. Supports profiling and matter-level access controls.
SharePoint / OneDrive
Microsoft Graph API. Indexes document libraries, team sites and SharePoint lists. Compatible with Microsoft 365 tenancies.
Outlook / Exchange
Email indexing via EWS or Microsoft Graph. Queries span documents and correspondence in a single search.
Aderant / Elite 3E
Matter metadata from your PMS is used to scope searches by matter, client and practice area.
Custom repositories
Any system exposing a REST API, CMIS endpoint or ODBC connection can be indexed. Includes legacy DMS platforms.

Integration architecture

iManage / SharePointDMS layerAria RAG Enginevector index + LLMPrecedent Libraryindexed on deployAnswer + Citationssource doc + page refFee-earner InterfaceDMS / email / chatAudit Logevery query storedOverride Flagescalation pathAria queries only your indexed content. No external data. Every answer is traceable to a source document.

Technical specification

Integration method
REST API connectors with OAuth 2.0 authentication. CMIS for legacy DMS platforms. Incremental indexing triggered by document events.
Indexing
Documents indexed on deploy and updated in near-real-time as content changes. Supports PDF, DOCX, MSG, XLSX and 40+ formats.
Retrieval model
Hybrid dense-sparse retrieval (BM25 + embeddings). Re-ranking layer tuned on your document corpus for precision improvement.
LLM layer
Anthropic Claude or Azure OpenAI deployed inside your cloud tenant. No data transmitted to external APIs unless explicitly authorised.
Response format
Answer text with inline citations. Each citation links to the source document and page reference. Confidence indicator per response.
Latency
Typical response time 1.5–4 seconds for queries across 500,000+ documents, depending on infrastructure.
Access control
Respects existing DMS permissions. A user can only retrieve documents their account has access to in the source system.

Regulatory compliance

SRA Code r.4.2Supervision requirement: every AI-assisted output is logged with source citations. Partner review checkpoints configurable per matter type.
ICO UK GDPRProcessing basis documented. Data residency within your own environment. No personal data transmitted externally.
SRA AMLClient and matter data used for retrieval only. No training on client data without explicit consent.

Discuss the integration with your DMS infrastructure

A technical conversation about your existing systems and integration approach. 30 minutes.

Book a technical call