Outbound Voice Agents (in development)

Building AI agents that make outbound calls and close the loop

We’re developing an intent‑first agent that confirms, reschedules, or routes calls for operational teams. Today we’re seeking design partners for a pilot and pre‑seed investors aligned with our thesis.

Request the investor briefApply to be a pilot site
LLM used for intent only•SignalWire / SIP compatible•HIPAA‑conscious design (in progress)

What the prototype does today

Simulates phone calls with a web UI
Opens with identity check (e.g., “Is this Chris?”)
Classifies replies (yes / no / wrong number / voicemail)
Emits structured outcomes via webhooks for downstream systems
Note: production features like full rescheduling flows and compliance reviews are on the roadmap below.
Thesis

Intent‑first voice automation beats script‑reading bots

Use LLMs to understand, not to improvise. Keep calls short, auditable, and reliable.

Why now

LLMs are excellent classifiers; unreliable storytellers.
Telephony APIs are mature; ops teams want measurable outcomes.
Compliance pressure favors deterministic flows.

What’s different

LLM for intent only; actions are explicit, logged, and reversible.
Modular platform: orchestration, telephony, secure data.
Builder‑first: SDKs and webhooks from day one.

Where this goes

Healthcare reminders → service bookings → collections follow‑ups.
From outbound only → mixed inbound/outbound play.
From pilots → verticalized products.
Roadmap

Near‑term milestones

What we’re shipping next to reach pilot‑ready.

Engineering

Initiates outbound calls via SignalWire (toll‑free supported)
Timeline/observability view for call outcomes
Reschedule flow: propose/confirm slots via webhook
Live transfer (warm) for escalations
Opt‑out & DNC list handling
Tenant‑aware data layer with encryption hooks

Go‑to‑Market

Pilot playbook & success metrics (answer, confirm, opt‑out rates)
Early‑adopter agreements (BAA addendum optional)
Investor brief & 10‑slide deck
Reference demo with synthetic data
Pilot

Design‑partner program

Limited slots for teams that can trial in a controlled, low‑risk setup.

What you get

Configured outbound flow for one use‑case
Weekly check‑ins & shared metrics
Priority on features needed for your workflows

What we ask

Small, representative call lists
Outcome data for evaluation
Permission to reference anonymized results

Ideal partners

Healthcare, services, or logistics teams
Existing calendars/CRMs with webhook access
Clear success criteria within 6–8 weeks
Apply for a pilot
Investors

Pre‑seed conversations

We’re raising to reach pilot proof and early revenue.

Use of funds

Engineer time to harden the orchestration & observability
Compliance & security review
Pilot support and integrations

Risks we’re de‑risking

LLM misclassification → confidence thresholds & fallbacks
Telephony edge cases → controlled pacing & retries
Compliance → intent‑only + audit trails

How to engage

Request the investor brief & demo recording
Intro call with founder
Tech deep dive (30–45 min)
Request the brief
FAQ

What people ask us at this stage

Do you use the LLM to speak for the agent?
No. We use the model to classify intent and pick actions. This keeps calls short and auditable.
What works today vs what’s next?
Today: outbound dialing via SignalWire, identity check, basic intent classification, and webhooks for outcomes. Next: rescheduling flows, transfers, richer analytics, and compliance reviews.
Is this HIPAA compliant?
We’re designing for HIPAA‑conscious deployments and plan to support BAAs for covered entities during pilots.