Skip to main content

FAQs (FAQ Generator)

The FAQs tab under Account Settings → System (/settings/faqs) holds the tenant-facing FAQ catalogue used by the ChatBot on WhatsApp and email. Operators can edit entries by hand or run the FAQ Generator — an AI scan of recent tenant conversations and IF/THEN rules that proposes new draft Q&A pairs.

First-time workspace setup

Configure ChatBot persona and IF/THEN rules first (setup step 11 companion), then run the FAQ Generator after you have real tenant threads. Pair with Getting Started — Recommended Setup Sequence. Deep link: platform.vivin.app/settings/faqs.

Direct access

Open FAQs at platform.vivin.app/settings/faqs. Route reference: Deep Links — Account Settings.

Pair with related settings

FAQ drafts feed the same tenant assistant as ChatBot persona / IF/THEN rules. Token cost meters as faq_generation on AI usage API. Contrast operator AI Chat (landlord_chat) — a different surface and ledger feature.

Who can edit FAQs​

The FAQs tab uses the same Chatbot permission gate as ChatBot settings. Roles without that permission do not see the tab.

FAQ list​

The main panel shows:

ControlMeaning
Run FAQ GeneratorOpens the cost-estimate modal, then scans history and appends draft FAQs
Last runTimestamp of the most recent generator run
Published countHow many FAQs are currently published vs total (for example 3/18 published)
Toggle on each rowPublish / unpublish — unpublished drafts stay internal until you turn them on
Edit / DeleteChange question and answer text, or remove an entry

Account Settings — FAQs tab with FAQ Generator header and draft/published FAQ list

Nothing publishes automatically — every generated entry lands as a draft until you flip its toggle.

Run FAQ Generator​

  1. Click Run FAQ Generator.
  2. Choose History to scan (1m–6m) and AI model (Sonnet 4.6 or Opus 4.8).
  3. Read the estimate: WhatsApp message count, email count, IF/THEN rule count, and approximate token / USD cost.
  4. Confirm Run FAQ Generator — or Cancel to abort without metering a run.

Account Settings — FAQ Generator modal with history window, model choice, and cost estimate

Walkthrough: open FAQs, Run FAQ Generator, review the estimate (3m / Sonnet), then Cancel without metering a paid run.
AI usage metering

Each confirmed run records faq_generation in the append-only ai_usage ledger (internal USD cost). Poll GET /ai-usage/summary?feature=faq_generation with a management JWT — see AI usage API — Feature values. There is no management-frontend ledger screen yet.

Review and publish​

After a run:

  1. Expand each new draft, edit question/answer wording for your brand voice.
  2. Publish only entries you want the tenant chatbot to use.
  3. Delete weak or duplicate suggestions.

Additive behaviour: new FAQs are appended and ordered from most to least common — existing published entries and your edits are kept.