Resources / Product Brief

Knowledge & Content, Product Brief.

FAQ management, knowledge base authoring, content versioning, multi-language support, and the AI training layer that ensures every agent response is accurate and on-brand.

Book a demo See Support →

What Knowledge & Content does

Ephanti Knowledge & Content is the layer that keeps AI agents accurate. It lets your team author, review, version, and publish the content that MEVA draws on when answering questions.

Unlike a static help centre, the knowledge base is the live input to the AI reasoning layer. When an article is updated, the change is reflected in agent responses immediately, no model retraining required. The Knowledge Gap report surfaces questions the AI couldn’t answer confidently, giving your content team a prioritised backlog.

How it works

1. Author

Write FAQs, articles, and policies in the editor. Add variants by language, channel, or customer tier.

2. Review

Route for approval. Publish when verified. Previous versions retained for rollback.

3. AI draws on it

MEVA retrieves relevant content at query time, assesses confidence, and cites the source in the agent panel.

4. Gap report

Questions the AI couldn’t answer confidently surface in the Knowledge Gap report for the content team to fill.

Key capabilities

Article editor: rich text, structured FAQs, policy documents, and SOPs
Multi-language support: one article, multiple language variants, auto-suggest translation
Version control: full history, diff view, one-click rollback
Approval workflows: draft, review, publish states with role-based permissions
Article scoping: which articles are available per channel, customer tier, or agent type
AI confidence scoring, see how confidently the AI draws on each article
Knowledge Gap report: questions the AI could not answer, grouped by topic and volume
Source citation in agent panel, agents see which article informed the AI suggestion
Search across all articles with relevance ranking
Content performance: how often each article is used, by channel and agent

Configuration options

Approval workflow: who can publish, who must review
Article categories and tagging structure
Language variants, which languages are required per article type
Confidence threshold: below X%, flag for human review rather than auto-answer
Content scoping rules: article availability by channel, tier, or industry
Review cycle reminders, auto-flag articles older than N days for re-review

Tier availability

Capability	Starter	Pro	Enterprise
Knowledge base authoring	✓	✓	✓
Multi-language variants	✓	✓	✓
Version control	,	✓	✓
Approval workflows	,	✓	✓
Knowledge Gap report	,	✓	✓
Article scoping by channel/tier	,	,	✓
AI confidence scoring per article	,	,	✓

FAQs

How does Ephanti use the knowledge base to answer questions? ⌄

When an agent encounters a question, MEVA searches the knowledge base for relevant content, assesses confidence against a configured threshold, and either uses the content to respond directly or flags low-confidence answers for human review.

Can I control which articles the AI uses? ⌄

Yes. Articles can be scoped to specific channels, languages, customer tiers, or agent types. You can also mark articles as draft (not available to AI) or verified (available), and set expiry dates for time-sensitive content.

How do I know when the knowledge base is out of date? ⌄

The Knowledge Gap report shows questions the AI could not confidently answer, grouped by topic and sorted by volume. This becomes the prioritised content backlog for your knowledge team.

See Knowledge & Content in action.

Book a demo and we’ll walk through knowledge base setup for your industry and use cases.

Book a demo All product briefs →