Metadata Tagging
Definition
Metadata is structured information about a piece of content rather than the content itself. For knowledge base articles, metadata might include: category (billing, integrations, security), product area, target audience (admin, end-user, developer), language/locale, publication date, last-reviewed date, author, and custom tags. This metadata serves multiple purposes: it powers faceted navigation in the help center, enables filtered search (show only billing articles), allows AI systems to retrieve content appropriate to the user's context (only return articles in the user's language), and supports content governance (identify articles not reviewed in the past 6 months).
Why It Matters
Metadata tagging transforms a flat list of articles into a structured, navigable knowledge system. Without metadata, search must rely solely on article text — returning everything that matches regardless of relevance to the user's specific context. With rich metadata, the AI retrieval system can scope its search to the most relevant subset: articles in the right category, for the right product version, in the user's language. This precision significantly improves answer accuracy and relevance.
How It Works
Metadata is added to articles manually by authors at creation time, or automatically during ingestion using classification models that analyze content and assign categories. At retrieval time, metadata filters are applied before or after vector search: pre-filtering restricts the search space to matching documents; post-filtering re-ranks results by metadata relevance. In practice, a combination of both is common.
Article Metadata Schema
How to Connect Salesforce to Your Knowledge Base
Tags
Real-World Example
An AI chatbot serves both free and paid users. When a paid user asks about an advanced feature, metadata tagging ensures the AI retrieves only articles tagged 'audience: paid-tier' rather than also returning free-tier articles about similar topics. The answer is precisely tailored to what the user has access to, avoiding confusion about features they cannot use.
Common Mistakes
- ✕Applying inconsistent metadata — tags are only useful if applied uniformly across all articles.
- ✕Over-tagging articles with too many categories, making filtering less meaningful.
- ✕Not including date metadata — without last-reviewed dates, there is no way to systematically identify and update outdated content.
Related Terms
Knowledge Base Article
A knowledge base article is a single piece of content within a knowledge base — covering one topic, question, or procedure in depth. Articles are the atomic unit of a knowledge base, and their quality, structure, and searchability directly determine how useful the knowledge base is for both human readers and AI retrieval systems.
Content Taxonomy
A content taxonomy is the hierarchical classification system that organizes knowledge base articles into categories and subcategories. A well-designed taxonomy makes content easy to browse and navigate, improves search filtering, and helps both humans and AI systems understand the scope and context of individual articles.
Knowledge Base Search
Knowledge base search is the capability that enables users to find relevant articles, and enables AI systems to retrieve relevant content to answer questions. Effective search combines full-text keyword matching with semantic understanding — finding relevant content even when users use different words than those in the articles.
Content Freshness
Content freshness refers to how current and up-to-date knowledge base articles are. Fresh content produces accurate AI answers; stale content produces confidently wrong answers. Maintaining freshness requires review workflows, expiry policies, and systematic audits that keep articles aligned with the current state of the product.
Ready to build your AI chatbot?
Put these concepts into practice with 99helpers — no code required.
Start free trial →