PlatformAI Voice Calls
It calls. You read four lines.
Two-minute screens, recorded and summarised, with live transfer to your desk.
Kandi · AI Voice Calls
How it runs
Call
Your questions, your script, the candidate's timezone.
Record
Audio and transcript stored for every call.
Summarise
Outcome, next step, availability, interest level.
Hand over
A keen candidate transfers live to your phone, mid-call.
Built in
Provenance-gated
Numbers bought from data vendors are refused for AI calls.
Voicemail discipline
A different short message per attempt, or silence and retry.
Quality-reviewed
Every call is scored after it ends, so drift surfaces early.
Questions buyers ask
What happens if the candidate says no?
The agent thanks them and ends. The outcome is recorded, the sequence stops, and they aren't called again for that role.
Is SMS available too?
SMS and WhatsApp ride the same sequences with the same opt-out rules. The held-before-send content checks are an email thing today.
What about voicemail?
The agent leaves a short message naming the recruiter and the role — a different one on each attempt, never the same twice — or stays silent and retries, depending on configuration. Dead and full mailboxes are detected and skipped.
Are recordings kept?
Every call's audio and transcript are stored and listenable, and a quality reviewer scores each call after it ends, so a script that's going wrong surfaces early rather than after a hundred dials.
Is this compliant to run in the US?
Numbers are screened by provenance before dialling — the policy refuses AI calls to numbers bought from data vendors and fails closed on unknown provenance. Calling windows run per candidate timezone, opt-outs are organisation-wide, and outbound numbers are registered for the SHAKEN/STIR attestation US carriers check. See Control.
See it on one of your own roles.
Bring a clinical role. See how the platform handles sourcing, outreach and initial screening for your team.