Knowledge Maze — AI Knowledge Base, Help Centre & Support Answers (KMS)
od: Maze Worldwide Limited
AI knowledge base, help centre, cited answers, articles, guides, workflows, translation, analytics
Knowledge Maze is an AI-native knowledge management system that unifies a governed knowledge repository, cited AI answers, decision-tree guides and omni-channel publishing in one platform. It is knowledge base software for organisations that answer questions at scale: one system for articles, approvals, translations, channels and answer quality, with an AI assistant doing the searching and drafting that consumes a support team's week.
Who it's for
- Customer service and contact-centre leaders who need agents to give fast, consistent, correct answers.
- CSR and support agents who need one place to ask a question and get a cited answer.
- Knowledge managers and content authors running review and approval workflows across a growing article library.
- Legal, compliance and governance reviewers who need approved content, version history and an audit trail.
- Regional and business-unit owners who need local content on top of a shared core knowledge base.
- Digital and self-service teams publishing the same approved answers to website FAQs, chatbots, mobile apps and email.
- Support operations leaders measuring deflection, content gaps and AI model spend.
The problem it solves Support teams run on knowledge nobody has time to keep right. Answers live in documents, wikis and inboxes that drift out of date, so agents give different answers to the same question. Articles are published without a clear review step, so nobody can say which version is approved. Regional teams copy and edit content, and the copies diverge. Translations lag behind the source. Chatbots and website FAQs are updated separately from the agent console, so customers and agents hear different things. When an agent can't find an answer, the gap is passed along by chat or email and never reaches the people who could write it.
Separate wikis, chatbot tools and help centres make this worse: none share a common, governed source of truth. The result is agents searching instead of helping, inconsistent answers, and content decisions made without data.
What Knowledge Maze does
- Ask Mazie, the AI assistant. A chat-first assistant that answers from your approved articles, with numbered citations, an Insights summary and Next-steps checklist. It can offer two answer variants side by side, and the agent's choice is recorded as feedback to improve the model.
- Governed knowledge repository. Articles move through Draft → In review → Approved → Published, with verification badges and full version history, so everyone can see what is approved.
- Business Units. A Core knowledge base plus regional business units, with Core content visible read-only inside each unit and roles assigned per unit.
- Omni-channel publishing. Control per article whether it appears on the website FAQ, website chatbot, CSR console, mobile app or email, and preview it in realistic channel frames in the Playground before it goes live.
- Translation. Translate articles per language, with per-language approval before publishing.
- Content gap management. Agents raise a gap from chat or search. It is routed to an inbox for content authors (New → Assigned → Drafting → Resolved) with a prefilled draft.
- Decision-tree guides. Step-by-step guides with a player and an editor, for processes that need a path, not a paragraph.
- Analytics. Usage, deflection and model-spend reporting to show what is working and what is missing.
- Admin console. AI model registry and routing rules, budgets, users and role-based access, governance guardrails and audit, localisation, appearance and a no-code Studio.
AI that does the work
- Ask Mazie reads your approved knowledge and answers with sources, so agents spend time resolving cases instead of searching.
- Model choice. Route chat to the model that fits the job and the budget, built on Microsoft Azure OpenAI Service.
Works with the systems you already run
Knowledge Maze complements your existing helpdesk, CRM and channels, publishing approved knowledge where customers and agents already work.
Getting started
Knowledge Maze is delivered as a fully managed, browser-based SaaS: nothing to install and no infrastructure to run. A live evaluation environment lets you try it with sample tenants and content first.
Find more details at https://maze.global/products/knowmaze