01
Your knowledge
Docs, FAQs, PDFs, policies, URLs. Chunked, embedded, versioned, and scoped to your business.
How it worksOne shared platform.
One shared platform.
Every business fully isolated.
Every business on Luran operates in its own sealed environment with its own knowledge, configuration, conversations, and customers. The infrastructure underneath is shared. The data never is.
Pipeline
Grounded retrieval, every time.
Below the confidence threshold, Luran asks a follow-up or hands off. It will not invent an answer.
02
Retrieval
The agent only searches your business's index. Below the confidence threshold it asks back or escalates. It never guesses.
03
Grounded reply
Replies compose inside strict system and business prompts. Every claim is cited back to source.
The pipeline ensures every answer:
Comes from your knowledgeFalls back when unsureCites source on demandFollows each business's own rules
Flow runtime
The same engine resolves a refund and onboards a customer.
Flows chain inputs, branches, API calls, knowledge lookups, and human steps. Flows move through draft, active, paused, and archived states, with execution history and analytics. Build guided journeys your AI can run in any conversation.
- TriggerAI ended turn, message keyword, schedule, webhook
- ActionHTTP, lookup, send message, update profile, dispatch
- BranchScore, sentiment, intent, customer attributes
- HumanHand off to an agent or assign by skill
Setup
Live in four steps, not four weeks.
01
Create your workspace
Name, brand, domain. We provision a sealed environment with isolated indexes and config.
02
Upload knowledge
Docs, FAQs, PDFs, URLs. Auto-chunked, embedded, versioned. Visibility controls per file.
03
Configure flows
Pick from starter templates or build your own. Refund, onboarding, escalation, scheduling.
04
Ship & monitor
Drop the widget in. Watch conversations land in the inbox, with analytics from minute one.