From recording to finalized review
Campaign-native AI call QA
Upload a call. AI transcribes it, scores it against that campaign’s policy, and cites the transcript. QA verifies, coaches, and finalizes — in one workspace.
AI evaluation
Sales outreach · Policy matched
Evidence from transcript
This call may be recorded for quality…
Okay, go ahead.
How it works
The full AI path, call by call
This is the same sequence the product runs: upload audio, transcribe with speakers, evaluate against the campaign scorecard, ground every finding in the transcript, then let QA close the loop. Play it through, or click a step.
Upload calls
Drop MP3, WAV, M4A — or choose a file.
- 01 · Uploaded
Audio is attached to a campaign and an agent. The file is stored; processing is queued.
- 02 · Transcribing
Speech-to-text splits speakers and timestamps so evidence can land on a real line, not a summary.
- 03 · Evaluating
AI loads that campaign’s active policy and scores every criterion against the transcript.
- 04 · Needs review
Findings, overall score, and evidence spans are ready. The call sits in the QA queue.
- 05 · Finalized
A reviewer accepts or overrides, coaches the agent, and locks the score into reporting.
AI QA that follows your campaigns
Activate policies per campaign. Upload calls. Let AI score them — then let QA review with evidence, not a spreadsheet.
Transcribe
Call audio becomes a speaker-labeled transcript with timestamps — the substrate every later finding cites.
Evaluate
The campaign’s active scorecard grades greeting, disclosure, empathy, close, and the rest — automatically.
Evidence
Each finding quotes the exact utterance and timestamp so QA can verify, not guess.
Review
Override a result, leave coaching notes, finalize. Campaign and agent metrics update from that review.
Call center management, built in
Scoring only works if the workspace already knows the campaign, the agent, and the policy. CallQX is that workspace — not a scoring tool bolted on later.
- CampaignsThe unit of work. Every call, policy, and agent assignment hangs off a campaign — Solar, Med, Insurance, and so on.
- PoliciesUpload the scorecard you already use. Activate it for that campaign. AI evaluates against the document, not a generic rubric.
- AgentsCreate call agents, assign them to campaigns (and teams), and attribute every recording to the person who took it.
- TeamsGroup agents for ownership and reporting — team scores roll up from finalized reviews.
- QA usersInvite-only reviewers who own the queue: verify evidence, override findings, coach, and finalize.
Campaign-native by design
Calls, policies, agents, and reviews all hang off campaigns — so AI QA matches how your call center actually runs. A Solar call is never scored with a Med script.
AI scores against your policies
Disclosure, script, empathy, closing, and the rest — pass, partial, or fail, with an overall score driven by the active campaign scorecard. You upload the document; AI derives the criteria.
Evidence, not guesses
Reviewers click a finding and the matching transcript lines highlight and scroll into view. Optional: seek the recording to that moment. Coaching stays grounded in what was said.
Hear every call as data
Upload call audio (MP3, WAV, M4A, and more). Get a clean, speaker-labeled transcript ready for policy-aware scoring — no manual tagging, no spreadsheet scorecards.
Admins run the workspace. QA closes the loop.
Two portals on the same origin. Agents are people whose calls get scored — they don’t log in.
Admin portal
Admin
- Create campaigns, activate policies, manage agents and teams
- Invite QA reviewers
- Upload calls and watch the queue
- Dashboard: volume, scores, compliance, leaderboards
QA portal
QA
- Work the needs-review queue, filter by campaign
- Open a call: audio, transcript, findings, evidence
- Accept or override each finding with a reason
- Coach the agent, then finalize
From campaign setup to finalized review
Stand up the workspace once. After that, every recording follows the same path — campaign policy in, coached agent out.
- 01
Set up
Create campaigns, upload and activate scorecards, assign agents and teams, invite QA.
- 02
Upload
Drop call audio with campaign + agent. Processing starts: transcribe, then evaluate.
- 03
AI evaluates
Campaign policies score the transcript. Every finding ships with quoted evidence.
- 04
QA finalizes
Reviewers override if needed, coach the agent, lock the review. Metrics update.
Run QA the way your campaigns run
Create an org, set up campaigns and policies, invite QA — then upload your first call and watch it move from audio to a finalized review.