Premium analytics for
AI and human voice calls.
VoiceLens grades every call across your call-center — whether an AI agent or a human agent is on the line: fault detection, latency outliers, slot capture, sentiment shifts and ASR quality, surfaced in a dashboard your team actually wants to open.
Free to start · Works with any call platform · SOC 2-friendly architecture
Call · 11 May 2026
Laidio Realtors · inbound enquiry
Duration
4m 12s
Agent talk
62%
Avg latency
0.94s
Slots captured
9 / 11
Agent latency · 12 turns
Fault
Repeated greeting
Agent re-greeted at 02:48 after customer correction.
Fault
Dead air · 4.3s
Silence at 03:14 after ID confirmation request.
Why VoiceLens
Call quality, finally a first-class product.
The signals you need to grade a call — from an AI agent or a human one — live in the transcript. We turn them into something you can act on.
Fault detection that finds what manual QA misses
Repetition, context loss, missed corrections, dead air, unconfirmed slots — flagged with timestamps and evidence, for AI and human agents alike.
Latency outliers with thresholds you control
Per-turn response latency for bots and humans, clustered against P95/P99 baselines so slow turns and regressions surface instantly.
Slot capture, scored end-to-end
Track every entity the agent had to capture — name, address, time slot — and how many got volunteered or corrected.
ASR & diarization quality, per call
Word-level confidence and speaker confidence aggregated into a single quality grade per conversation.
Built for pipelines, not just demos
Upload a call recording or POST a transcript and get a graded analysis back — then fan it out to your QA queues with signed webhooks.
Premium dashboard your QA team will use
Dark, fast, keyboard-friendly. Every fault is one click away from the exact transcript turn it came from.
How it works
From transcript to insight in three steps.
Stream calls into VoiceLens
Upload a recording or drop a transcript into the dashboard — or POST it to the API after every call, whether a bot or a human agent handled it.
Get a graded analysis in seconds
Faults, latency outliers, slot capture and sentiment surfaced automatically — no rules to write, same scorecard for every agent.
Act on the calls that need it
Send flagged calls to your QA queue, pin patterns, and turn them into better prompts for your bots and better coaching for your agents.
Made for
Call-center teams — AI agents, human agents, or both.
Whether you're scaling a sales dialer, coaching a floor of human agents or replacing tier-1 support with AI, VoiceLens gives you the QA layer your team has been hand-rolling in spreadsheets.
Outbound AI voice agents
- Detect dead air & repetition
- Score slot capture per attempt
- Compare agent versions
Inbound support automation
- Catch missed pivots
- Track resolution rate
- Flag empathy gaps
Human agent teams
- Grade every call, not a sample
- Coach with timestamped evidence
- Benchmark agents against bots
Quality + compliance teams
- Severity-graded fault feed
- Evidence pinned to transcript
- Exportable per-call reports
Pricing
Simple, usage-based pricing.
Start free. Upgrade when you outgrow it. Talk to us when you need SOC 2.
Starter
Free
For teams grading their first calls
- 300 call minutes / month, audio or transcript
- 1 workspace with roles & team invites
- Fault, latency, slot & sentiment analytics + AI summaries
- Rubric templates, case branching & golden-set calibration
- 3 AI rubric drafts / month
- Alerts, weekly digest, signed webhooks & API keys
- 1 embedded dashboard per workspace & shareable scorecards
Pro
$0.05 / call
Pay only for what you grade
Everything in Starter, plus
- No monthly minute cap, pay as you go
- Up to 5 workspaces
- 50 AI rubric drafts / month
- 10 embedded dashboards per workspace
- VoiceLens branding removed from embeds & scorecards
Enterprise
Talk to us
For regulated call-centers
Everything in Pro, plus
- Up to 50 workspaces
- Unlimited AI rubric drafts
- 100 embedded dashboards per workspace
- Committed volume with invoiced billing
- Custom allowances for your account
- SLA & private deployment options
Stop guessing how your agents are performing — AI or human.
Spin up a workspace in under a minute. Drop in a recording or transcript. See what your QA team has been missing.