Quackback splits AI into two groups: Basic AI ships on every tier with no usage caps, Advanced AI is included on Pro and Scale.
Basic AI (every tier)
- Semantic search across feedback — find related posts even when the wording differs, powered by vector embeddings on every post.
- Similar-post suggestions when filing — when admins triage incoming feedback, Quackback surfaces existing posts that look related so duplicates get merged early.
Advanced AI (Pro and Scale)
- AI post summaries — multi-paragraph summaries that capture the core user need, key quotes, and suggested next steps for each post.
- AI duplicate / merge suggestions — embeddings prefilter + LLM verification proactively flag true duplicates as they arrive, instead of waiting for you to search.
- AI sentiment analysis — every signal gets a positive / neutral / negative classification, exposed in trend charts and breakdowns.
- AI feedback extraction — pull individual feature requests and bug reports out of long-form sources (Slack threads, Intercom conversations, Zendesk tickets).
Usage limits
Pro and Scale include reasonable monthly usage limits on Advanced AI to keep things fair. We size the limits so most teams never hit them; if you approach the cap, we'll surface a warning in the dashboard and you can upgrade for more headroom. Basic AI is unlimited on every tier.
What doesn't count toward AI usage
- Embeddings used for semantic search, merge prefilter, and Help Center search. Cheap, pervasive, and always unlimited.
- Vector similarity search and full-text search.
- Failed retries — only the final successful attempt is billed.
- Pure SQL operations like analytics aggregations.
What happens at the cap
Advanced AI features pause until the start of the next billing month, or you can upgrade for more headroom. Existing AI-generated artifacts (summaries, sentiment scores) remain visible — only new AI work is paused.