COMPASS™
Knowledge Assistant

A simple, trusted platform for building enterprise-grade AI chatbots on your knowledge

What is it?

Secure and pre-built, enterprise grade AI chatbots

COMPASS™ is a production-ready knowledge retrieval (RAG) platform that enables organisations to deploy accurate, secure AI chatbots over their internal and customer-facing content.

It combines a pre-built AI engine, a configurable user interface, and plug-and-play connectors to enterprise data sources such as SharePoint, document repositories, APIs, and databases—without introducing new software licence costs.

Typical use cases include:

  • Customer support chatbots over product documentation, FAQs, and help content
  • Internal employee assistants for HR, IT, onboarding, and policy queries

Preview COMPASS™

Accelerated implementation

Why it accelerates chatbot implementation

Most chatbot projects stall due to complexity, tooling sprawl, and long build cycles. COMPASS™ removes that friction.

Key accelerators:

  • Pre-built core AI engine – no need to design RAG pipelines from scratch
  • Proven delivery model – discovery to production in 1-2 weeks depending on number of data sources and complexity
  • Open-source foundations – no licence fees, no vendor lock-in
  • Enterprise-ready from day one – security, access control, auditability, and scale built in
  • Configurable, not custom-heavy – adapt UI, connectors, and behaviour without rewriting the core

Simple by design

High-level architecture

COMPASS™ uses a clean, modular architecture built on trusted open-source tooling.

  • 1

    User interface (configurable)

    • Web-based chatbot UI or embedded into existing tools
    • Custom branding, conversation flows, and escalation rules
    • Optional admin and analytics views

  • 2

    Core AI engine (pre-built)

    • Retrieval-Augmented Generation (RAG) pipeline
    • Document ingestion, chunking, embeddings, and vector search
    • Prompting and retrieval logic tuned to minimise hallucinations
    • Supports commercial or open-source LLMs depending on requirements

  • 3

    Data connector layer (extensible)

    • Native connectors (e.g. SharePoint, document stores)
    • API-based connectors to any external system (CRM, ticketing, bespoke platforms)
    • Structured and unstructured data support
    • Permissions and access controls respected at source

  • 4

    Platform & operations

    • Runs on standard cloud or on-prem infrastructure
    • Monitoring, logging, and performance metrics
    • CI/CD and production-grade deployment patterns

What organisations get

  • A working, tested, production chatbot, not a prototype
  • Clear governance over data, access, and responses
  • A platform that can scale from 5 – 500+ concurrent users
  • Ongoing optimisation, monitoring, and support options

Implementation Plan

  • 1-2 weeks implementation project covering all aspects of architecture and user design, configuration, integration, testing and optimisation
  • Rollout training and support
  • Annual support plan
  • Self-managed environment or full-service options available

How does COMPASS™ compare

Comparison to others

COMPASS™ is the fastest, lowest-risk way to deploy a serious enterprise grade AI chatbot—built on open technology, designed for real knowledge, and ready for production.

Features

COMPASS™
Knowledge Assistant

Microsoft
Copilot

ChatGPT
(Enterprise / Teams)

Primary PurposePurpose-built platform for enterprise knowledge chatbots over internal & customer contentProductivity assistant embedded in Microsoft 365 ecosystemGeneral-purpose AI assistant with optional enterprise controls
Knowledge Retrieval (RAG)Native, pre-built RAG engine tuned to minimise hallucinationsLimited/native retrieval via Microsoft Graph & SharePointAvailable via custom GPTs or API, but not turnkey
Data Source ConnectivityPlug-and-play connectors (SharePoint, doc stores, APIs, databases)Strong for Microsoft sources; weaker outside ecosystemManual uploads, APIs, or custom integrations
Access Control & PermissionsRespects source system permissions nativelyInherits Microsoft 365 permissionsBasic workspace-level controls; custom work needed for fine-grained permissions
Deployment ModelCloud or on-premMicrosoft cloud onlyOpenAI cloud (Enterprise), limited on-prem options
Vendor Lock-inOpen-source foundations, no lock-inHigh lock-in to Microsoft ecosystemModerate lock-in to OpenAI platform
Licence CostsNo new AI software licence feesPer-user Copilot licences requiredPer-seat Enterprise/Teams pricing
Time to Production1-2 weeks to production chatbotFast to enable, slower for true knowledge botsFast to start, slow to productionise
Custom UI & BrandingFully configurable chatbot UI + embeddingLimited UI customisationLimited (unless building custom front-end)
Customer-Facing ChatbotsDesigned for external & internal usersPrimarily internal usersPossible, but requires custom build
Hallucination ControlsRAG + tuned retrieval & promptingGeneral-purpose AI responsesGeneral-purpose; depends on custom setup
Governance & AuditabilityBuilt-in enterprise governanceMicrosoft compliance toolingEnterprise controls, but less domain-specific
ScalabilityDesigned for 5–500+ concurrent usersScales with Microsoft tenancyScales, but knowledge layer is DIY
Typical Use CasesHR, IT, policies, product docs, support, onboardingEmail, meetings, docs, summarisationGeneral Q&A, brainstorming, light internal knowledge
Implementation ApproachDelivered platform + implementation projectSelf-serve enablementSelf-serve + custom development
Best ForOrganisations needing a serious production knowledge botMicrosoft-heavy orgs wanting productivity AITeams wanting flexible general AI, not a dedicated knowledge platform