Where does the 95% AI recovery rate come from?
It's a modeling assumption, not a measured result. The 5% gap stands in for callers who insist on a
human handoff, callers who hang up before completing intake, and edge cases like wrong-number or robocall
traffic that the agent drops. We treat 95% as recovered missed calls, then apply your stated close rate to
that pool. If you want a more conservative view, drop the recovery to 85% in your head and re-read the numbers.
I already have an IVR. Does this still apply?
Yes. An IVR is technically an "answered" call, but callers can drop off in the menu tree before any
human or system completes intake. If your phone system shows a high answered-call count but a low
booked-call count, check your IVR abandon rate first. The AI voice agent replaces the menu maze with a
single, natural conversation.
How do you split after-hours leak from in-hours leak?
The heatmap above uses a modeled weekly inbound pattern for your industry, not measured call data.
We assume your front desk covers a typical 9-to-6 window, so the cells outside that window receive the bulk
of the missed weight. Your own call log will show your real pattern. You can override the miss
rate slider to model a tighter or looser coverage window. The math holds either way.
Can I keep my receptionist and just run the AI for overflow?
Yes, that's a common hybrid setup. The AI sits on a "press one for booking, press two for
existing patient, press three to reach a human" tree, or it routes by detected intent. Your team handles
the calls that genuinely need a human (insurance disputes, complaints, complex changes), and the agent
handles intake, scheduling, basic FAQ, and after-hours coverage. Net effect: your receptionist gets time
back for focused work, and after-hours callers reach someone who can book them.
Does the AI voice agent count as a robocall under FCC rules?
Answering is generally not. The TCPA restricts calls a business makes using an artificial or
prerecorded voice, and the FCC's February 2024 ruling on AI voices applies to AI technology that initiates
outbound calls. When a customer dials you and the AI answers, the customer started the call. The line flips
when the AI dials out: an outbound AI call, including a callback to a missed call, needs the person's prior
consent (written consent if it markets anything). Any outbound SMS goes through A2P 10DLC-registered
campaigns with documented consent on each opt-in. The privacy policy and SMS terms are linked in the
footer. This is general information, not legal advice.
What is in the detailed PDF report?
Five pages. (1) Your numbers: monthly leak, annual leak, hour-by-hour heatmap, and engagement payback. (2) The model defaults for your industry side-by-side with your inputs. (3) The exact intake script we would configure for your
category, with qualifying questions tuned to your average ticket. (4) A four-week implementation roadmap.
(5) An ROI scenario set for three plausible business outcomes. No marketing fluff. You receive it after
you submit, with a single confirmation SMS when the report is ready.