Build an LLM Wiki for each document or domain, then grow it into a knowledge infrastructure.
Concierge Builder organizes product materials, manuals, FAQs, policies, service explanations, and other documents into domain-specific LLM Wikis. Each Wiki can be tested in Test Mode and then published as a customer-facing or internal concierge on a website, help center, or company portal.
This is not a low-cost commodity chat widget. Concierge Builder is a public and testable knowledge layer for building supervised, official knowledge one domain at a time. Multiple LLM Wikis can later be integrated and routed through ConceptMiner’s GNG+MST conceptual structure model.
Build and Test Mode are free. Production use requires a monthly plan. LLM token costs are separate, and BYOK — Bring Your Own API Key — is recommended.
You do not need to build one massive knowledge base from the beginning.
Enterprise knowledge is naturally divided by product, department, process, version, customer-support domain, and document type. Trying to build a single company-wide RAG index from the start often creates problems before the pilot even proves value: chunk design, retrieval noise, unclear supervision responsibility, and update management.
ThinkNavi takes a different approach: build an LLM Wiki for each document or domain first, then integrate them later through ConceptMiner. Concierge Builder acts as the publication and validation layer for one domain-specific LLM Wiki.
For teams that operate customer support, internal knowledge, and product documentation.
Concierge Builder is not positioned in the low-cost chatbot market. It is designed for teams that care about accuracy, supervision, document quality, and long-term knowledge operations.
Customer support and help centers
Convert product documentation, FAQs, terms, and support policies into LLM Wiki articles, then operate help content and chat guidance as one system.
Internal knowledge and training
Organize internal rules, manuals, training materials, and department-specific documents as domain-specific Wikis for employee Q&A and onboarding.
Multiple products and document sets
Manage Product A, Product B, support policies, implementation manuals, and other document groups separately, then extend them into a cross-domain knowledge base when needed.
Consultants and SIs designing knowledge systems
Use Concierge Builder as a tool for designing and supervising customer knowledge infrastructure. It is not for reselling a chat widget alone, but for supporting knowledge operations.
Start with one domain. Test it. Publish it. Integrate later.
Create one domain-specific Wiki from a URL, file, or template. Review the answers before publication, then publish it as a concierge. When multiple domains need to be connected, you can extend the system through Knowledge Base Builder and ConceptMiner.
Templates are starting drafts for organizations whose documents are not yet ready.
Concierge Builder can generate an AI concierge from URLs and files, but it can also start from industry templates. Templates are not the final knowledge base. They are starting drafts for organizations that do not yet have complete FAQs, help articles, or manuals.
In production use, the concierge should be improved by adding each organization’s official documents, FAQs, help articles, manuals, and supervised explanations. Industry templates will be expanded gradually based on user requests.
SaaS / product support
Pricing, features, onboarding steps, troubleshooting, support escalation, and account-related guidance.
Manufacturing and technical support
Specifications, supported materials, equipment, maintenance, safety information, and pre-estimate interviews.
Professional services
Consultation areas, procedures, required documents, consultation flow, retainer contracts, and client guidance.
Clinics and medical offices
Office hours, reservation methods, first-visit guidance, required items, access, and frequently asked questions.
Construction and renovation
Service scope, consultation-to-estimate flow, project examples, warranty information, and common concerns.
More templates on request
Please contact us if you need a template for a specific industry. Templates will be added and improved gradually based on real use cases.
The purpose is different from low-cost embedded chat widgets.
Concierge Builder is not a low-cost widget for displaying a single FAQ. Its value is not the chat UI itself, but the combination of readable LLM Wiki articles, supervised official knowledge, domain-specific publication points, and future cross-domain integration.
For that reason, it should be evaluated not as a monthly fee for a chat widget, but as a platform for publishing, validating, and operating knowledge bases.
Build and test for free. Pay monthly only from production use.
You can generate an LLM Wiki and test answer behavior for free. A monthly plan is required only when the concierge is published, embedded, or used in production.
Build & Test
Create Wikis, edit articles, use Test Mode, and validate the system before internal or customer-facing production.
- Create from URLs, files, or templates
- Generate and supervise LLM Wiki articles
- Check responses in Test Mode
- Validate before production publication
Account / Multi-domain Plan
A production account plan for operating multiple domain-specific Wikis, AI concierges, and publication points.
- Designed for multiple production domains
- Publish to websites, help centers, or internal portals
- Embeddable concierge widget
- BYOK recommended — LLM token costs are separate
- Credit purchase is available if an API key cannot be provided
※ Please check the pricing page for the number of production domains, usage conditions, annual plans, and other details. ※ LLM token costs are not included in the monthly plan. Using your own API key is recommended.
Concierge Builder and Knowledge Base Builder serve different roles.
They are part of the same product direction, but they target different stages. Concierge Builder is for publishing, validating, and embedding a domain-specific LLM Wiki. Knowledge Base Builder is for integrating multiple Wikis and internal documents into a more complete knowledge base.
| Item | Concierge Builder | Knowledge Base Builder |
|---|---|---|
| Main role | Publish, validate, and embed domain-specific LLM Wikis | Build integrated knowledge bases from multiple documents and Wikis |
| Typical scope | Product materials, FAQs, help articles, service explanations, one domain of documents | Internal documents, department knowledge, multiple products, company-wide knowledge |
| Publication style | Website embedding, help concierge, internal portal | Integrated KB, internal search, cross-Wiki answers, on-premises operation |
| Conceptual structure model | Used as a future integration layer | Used for integrating and routing multiple knowledge bases |
| Deployment mindset | Start with one domain and validate value | Integrate multiple domains into organizational knowledge infrastructure |
A knowledge foundation that can expand into ThinkNavi, ConceptMiner, and Concept Index.
An LLM Wiki created with Concierge Builder does not have to remain a standalone chat window. It can later connect to ThinkNavi’s integrated knowledge base, ConceptMiner’s GNG+MST conceptual modeling, and Concept Index for concept-based search on conventional relational databases.
ThinkNavi
The application layer for LLM Wiki, integrated knowledge bases, Concierge Builder, and thought-support workflows.
ConceptMiner
Builds conceptual structures across multiple Wikis and document groups using GNG+MST, forming the core of cross-domain routing.
Concept Index
A planned indexing layer that enables concept-based search on conventional relational databases such as PostgreSQL.
Start with one document or one domain, then grow your knowledge infrastructure.
Begin by creating an LLM Wiki from one document, one product, or one FAQ domain and test it for free. When multiple domains need to be connected, you can expand into Knowledge Base Builder and ConceptMiner.
Contact us for enterprise pilots, multi-domain operation, or early information about Concept Index.