AI PlatformProfessional ServicesRepresentative project

Enterprise RAG Knowledge Assistant

An internal knowledge assistant for Meridian Systems, grounded in years of case studies, proposals, and internal documentation.

Customer: Meridian Systems (demo)

Read the full case study

Meridian Systems, an operations consultancy, had years of institutional knowledge — client case studies, proposal templates, internal process documentation — spread across a document management system nobody could search effectively. We built an internal assistant that lets consultants ask questions in plain language and get sourced answers from that archive.

Challenge

New and existing consultants were repeatedly re-solving problems Meridian had already solved for previous clients, simply because there was no effective way to search years of accumulated case studies and internal documentation. Keyword search in their existing document system returned too many irrelevant results to be useful.

Solution

We built a permission-aware RAG assistant that indexes Meridian's internal document repository — case studies, proposals, methodology documents — and lets consultants query it conversationally. Results respect the same access controls as the source documents, so the assistant never surfaces content a given user couldn't already see directly. Every answer links back to the source document and page.

Architecture

Documents are ingested from Meridian's existing document management system via a scheduled sync, chunked, and embedded into a vector store. A metadata layer tracks the original access permissions for each chunk, enforced at query time so retrieval results are filtered per-user before reaching the language model. The assistant runs as an internal web application with authentication tied to Meridian's existing identity provider.

Key features

Permission-aware retrieval

Search results respect the same access controls as the source documents.

Conversational query interface

Consultants ask questions in plain language instead of guessing keywords.

Source-linked answers

Every response links back to the originating document and page.

Scheduled content sync

The knowledge base stays current with the document management system automatically.

Integrations

Document management system

Scheduled ingestion sync from Meridian's existing document repository.

Identity provider

Authentication and permission mapping tied to existing user accounts.

Representative outcomes

Faster access to prior work

Consultants can find relevant past case studies in seconds instead of asking around.

Reduced duplicated effort

Less re-solving of problems the firm had already worked through before.

Faster onboarding

New consultants can query institutional knowledge directly instead of relying on tenure.

Technologies

TypeScriptPythonPostgreSQLOpenAI

Have a similar project in mind?

We're glad to talk through the approach, even before you're ready to commit.