◆ Solutions · Sales / SDR
AI reps with a real face.
Outbound and qualification have collapsed into AI conversations on most teams. The blocker now is trust: prospects engage longer with a face than with a voice. NORTH is the face layer.
◇ Problem
Why current solutions fall short.
- Voice-only AI SDRs lose engagement after the first 60 seconds
- Existing avatar studios produce static videos, not adaptive conversations
- Per-seat avatar licensing doesn't fit usage-based outbound load
- Live demo calls require an SDR present; AI demos drift off-script
Stack-level cost: every blown demo is a real human-minute and a delayed pipeline opportunity.
◆ How NORTH fits
An AI SDR with the trust profile of a real video call.
- Run inbound 'book a demo' through a video qualification call
- Outbound dial-and-pitch with a live AI rep on screen
- Live product walkthrough: avatar narrates while sharing screen
- Group demo / webinar with one GPU stream, many viewers
- Hand off to a human in the same Zoom call when the lead converts
◆ Why now
Now: prospects accept AI sellers, but only if they look and feel real on video.
- Per-second billing matches volatile outbound volume
- Live demo mode: avatar walks the prospect through your product
- Multi-viewer mode: one render stream into many viewers (group demo, webinar)
- Bring your own LLM, voice clone, and CRM tools
- Same-day handoff to a human via warm Zoom / Meet bot transfer
◆ Integration shape
What it actually looks like in your codebase.
01
Pair with your CRM
Trigger sessions from your existing dialer / scheduler / inbound router.
02
Wire your stack
Your STT + LLM + TTS power the conversation. NORTH is the visual layer.
03
Pick the rep
One avatar per persona, or a roster mapped to industry vertical.
04
Ship outbound
WebRTC into the browser, or a Zoom / Meet bot for live calls.
◆ What to measure
The metrics that prove this is working.
Median session length vs voice-only
Compare against your existing voice-AI baseline; track time-on-call as a proxy for engagement.
Qualification rate
Video-first AI flows tend to convert higher on inbound; measure against a voice or chat control.
Render cost / minute
$7/hr prorated to the second. Outbound spend lines up with actual talk-time.
We don't publish customer numbers. The right comparison is against your existing baseline, your STT, your LLM, your voice, your traffic.
◇ FAQ for this use case
The questions that actually come up.
Will the prospect think it's a real person?
We require AI disclosure in our Terms of Service. The avatar is realistic enough to hold attention; the goal is engaged, transparent AI, not deception.
Can we use a real employee's likeness?
Yes, with documented consent from the employee. Public-figure / unauthorized-likeness use is prohibited under our terms. See /safety for the full policy.
How do we handle warm handoff?
Use the Zoom / Google Meet integration. The avatar joins the meeting via a standard meeting-bot framework; a human can join the same call and the avatar drops out.
What about call recording / compliance?
Recording happens in your stack, not ours. We don't retain WebRTC frames. Disclosure language and consent capture sit on the customer side, same as a human-led call.
Ready to ship?
Email if you want async, book if you want a structured call. We reply within one business day.