Skip to main content
Analyze completed calls with structured insights across language and performance. Pick a pack, choose a date range, and review average score, resolution rate, trends, and top user questions.

Watch the walkthrough

The English interface uses demonstration data. Runs, customers and results shown are simulated; use the subtitle menu for your language. Audio-level metrics (WER, overlapping speech) are not included in v1 — analysis is transcript-based. The judge reads the transcript together with the tools the assistant ran and their results. A metric that does not apply to a call, such as Tool-call accuracy in a call without any tool, stays empty instead of counting as 0.

QA packs

QA credits

QA runs consume workspace credits based on LLM usage for each analysed call. The Usage page shows charges as QA usage. Cost breakdowns and markup are not exposed in the product UI. Calls without caller speech are skipped and cost nothing.

Calls without caller speech

A run judges only calls in which the caller said something. A call where the caller never spoke, for example because they hung up silently after the greeting, is not judged, not charged, and not counted in the call total, the averages or the failed results. The report shows how many calls were skipped and lists each one as No caller speech. If the caller never spoke in any call of the selected range, the run does not start, and the API answers 400 with No calls with caller speech found in this range.

Availability

Requires AI Quality Assurance in your plan. This is distinct from per-assistant AI QA scorecards.

UI

Open AI QA in the sidebar (/qa). Create a run, wait for the analysis progress, then open the report on its Overview tab.
AI Quality Assurance cohort list and Create Cohort button

AI QA → Create Cohort starts a new analysis; open a completed cohort from this list to review it.

API

  • GET/POST /api/v1/qa/runs
  • GET /api/v1/qa/runs/{id}?include_results=true
Scopes: calls:read / calls:write. In the results, error: "no_caller_speech" marks a skipped call, and aggregate.skipped_calls.no_caller_speech counts these calls.

MCP

  • list_qa_packs
  • list_qa_runs
  • get_qa_run
  • create_qa_run

CLI

  • famulor list-qa-runs
  • famulor create-qa-run --pack full --from-at 2026-10-01T00:00:00Z --to-at 2026-10-28T00:00:00Z
  • famulor get-qa-run <id> --include-results true