AI Platform · US Technical Guide

Understand the Components Behind LuMay's AI Platform

A technical map for US teams evaluating LuMay component by component - what each layer is responsible for, where it hands off to the next one, and what to verify before you rely on it in production.

Overview

This page exists for one job: helping a US-based technical evaluator decide which part of LuMay's AI platform to read about next, and in what order. It is not a pricing or business-case page - that overview lives at /platform. Everything below is a map of the same eight technical components documented in detail on their own pages, with the specific question each one answers.

If you're starting from zero, read Architecture first - it defines the six layers that every other page below refers back to. If you already know which layer you care about (reasoning, workflow, integration, voice, deployment, security, or analytics), jump straight to that page.

Component Map

ComponentLayerWhat it answers
ArchitectureLayers 01-06The six-layer reference model and the contracts between layers - start here if you're mapping LuMay onto your own systems.
Core EngineLayer 02 · IntelligenceThe reasoning runtime every agent runs on - provider-agnostic LLM calls, RAG retrieval, structured outputs, tool dispatch.
Orchestration EngineLayer 03 · OrchestrationWhere workflow state, approvals, retries, and multi-agent coordination live, separate from the reasoning runtime itself.
Connectors & AdaptersLayer 04 · IntegrationWhat agents can read, write, and dispatch to in your systems of record, and how credentials and failures are handled.
Voice EngineLayer 01 · Interface (voice)The real-time speech pipeline - transcription, reasoning, mid-call tool calls, and synthesis - for phone-based agents.
DeploymentInfrastructureSaaS, customer cloud, private cloud, or on-prem Kubernetes - who manages what, and where data lives in each model.
Security & GovernanceLayer 05 · GovernanceAuthentication, RBAC, tenant isolation, PII controls, and audit trails - what's implemented versus what you configure.
Analytics & ROILayer 06 · AnalyticsBuilt-in KPIs, the observability stack behind them, and which metrics still require a separate survey integration.

Technical Evaluation Checklist

Before a deeper architecture review, most technical teams want answers to the following - each maps to one of the pages above:

  • Ownership boundaries - which layer is responsible for a given failure, and who do you escalate to? See Architecture.
  • Model dependency - what happens if you change LLM providers or a provider has an outage? See Core Engine.
  • Workflow auditability - how are approvals, retries, and version history recorded? See Orchestration Engine.
  • Integration scope - exactly what can an agent read or write in your systems, and how are credentials validated? See Connectors & Adapters.
  • Voice-specific dependencies- telephony, STT/TTS providers, and handoff behavior, if you're evaluating a phone-based agent. See Voice Engine.
  • Deployment responsibility - who manages infrastructure, patching, and recovery in each deployment model? See Deployment.
  • Security posture- what's enforced by the platform versus what you must configure yourself. See Security & Governance.
  • Measurement baseline- what's measured from day one versus what requires an additional integration. See Analytics & ROI.

Go Deeper

Each component page below includes the same structure: an overview, a request or request-flow walkthrough, a feature or responsibility table, and links back to the adjacent components it hands off to or depends on.

FAQ

Is this the same as the /platform page?

No - /platform is the business overview (four engines, pricing posture, trust pillars). This page is the technical component map used for architecture evaluation.

Which page should I read first?

Start with Architecture - every other component page refers back to the six layers it defines.

Do all eight components ship together, or can we adopt them separately?

Each component page describes its own responsibilities and dependencies; confirm directly which combination fits your first pilot rather than assuming all eight are required on day one.

Is there a US-specific hosting or data-residency guarantee on this page?

No - hosting and residency specifics are covered, with explicit limits on what's promised, on Deployment.

Who typically reviews these pages?

Solutions architects, security/compliance reviewers, and engineering leads scoping a pilot - see the checklist above for the specific questions each role tends to ask.

Ready to walk through this against your own stack?

Book a 30-minute architecture review. We'll go component by component against your systems of record, your deployment constraints, and your security requirements.