EnrollVoice — 24/7 AI admissions agent for small private K-12

1. One-liner

24/7 AI agent that answers small private school inquiries in 60 seconds and books campus tours before families go cold.

2. Trend signal — why now?

Three things lined up in the last 12 months:

1. ESA / school-choice tailwind. Texas launches the largest state school-choice program ever in Fall 2026 (~20K new Catholic seats alone). Arizona’s universal ESA hit 98,244 students by Dec-2025. New Hampshire: 11 of 28 Christian schools either newly opened or grew ≥50% in four years of vouchers. Trump admin pushed national voucher expansion. Demand-side faucet just got opened.
2. Voice-AI got cheap and good in 2025. Retell AI, Bland, Vapi, Synthflow — sub-300ms latency, FERPA/SOC-2 path, $0.05–$0.10/minute economics. K-12-tuned voice agents that actually book onto a Google Calendar without screwing up didn’t exist 12 months ago.
3. The math is brutally clear and getting cited. 5-min response is 100× more likely to connect than 30-min. 60-sec response = +391% conversion. Average school responds in 24–72 hrs. 70% of inquiring families drop off before applying due to slow/weak follow-up. Most students enroll at the first school that replies.

Schools know the problem. They can’t afford a 24/7 admissions human. The standard CRMs (Finalsite Apply, Blackbaud, Veracross, Ravenna) are priced and built for big-prep schools — not the 87% of US private schools that have <300 students.

Provenance:

• Signal 1: Speed-to-lead: 60-sec reply = +391% conversion; most students enroll at first responding school; 70% drop-off before applying — Reshape / GrowthJockey — May 2026
• Signal 2: Voice-AI agents now FERPA/SOC-2-compliant and sub-second-latency for K-12 (Retell, Bland, Vapi, Synthflow) — Retell AI / CloudTalk — Apr 2026
• Signal 3: ESA/voucher boom: TX launches biggest-ever ESA Fall 2026 (~20K Catholic seats); AZ at 98,244 ESA students; NH Christian schools grew ≥50% — EdWeek / NHPR — Feb–Apr 2026 Category: Tech-unlock

3. The opportunity

The market is a barbell. On the top end: Finalsite, Blackbaud, Veracross, Ravenna — purpose-built admissions CRMs costing $8K–$40K/yr, optimized for prep schools with 600+ students and a 3-person admissions office. On the bottom end: SchoolCues at ~$1/mo, Skooly, Newton (KaiPod) — generic SIS modules with admissions as a stepchild feature. Neither end actually answers a parent who emails at 9pm Tuesday. Neither calls back the 4pm Friday voicemail before the family tours a competitor on Saturday.

The wedge is a focused product that does one thing: takes a brand-new inquiry through to a confirmed campus tour, in any channel the parent uses, faster than any competitor. It bolts on top of whatever SIS/CRM/email the school already uses. It doesn’t try to be the system of record.

Disrupted incumbents:

• Generic AI voice platforms (Retell, Bland) — schools can’t configure them, don’t know what to ask, don’t know what good follow-up looks like.
• Big admissions CRMs — too expensive, too long to deploy, and ironically don’t solve speed-to-lead.
• Free school CRMs — no AI layer, no after-hours response, no drop-off radar.

What an AI-first focused tool does 10× better: every inquiry gets a specific, school-voice response within a minute, every tour has a pre-brief sent the night before, every family that goes quiet for 5 days gets a personalized re-engagement, every head-of-school sees a Monday-morning “hot families this week” digest. That’s the loop schools can’t run today.

4. Target market

• Primary customer: Small private K-12 admissions director (or, for the smallest schools, the head-of-school doing admissions on the side) at:
  • Christian / classical / Catholic / Montessori / Waldorf schools, 50–300 students, US
  • Tuition $8K–$22K/year per family
  • 1 person handles inquiry → tour → app → enrollment
  • Member of ACSI, NAIS regional, NCEA, ICLE, or unaffiliated
• Secondary: Microschool founders (10–60 students), often a former teacher running on Prenda, KaiPod, Acton, Wildflower, or independent. ~95K microschools in the US, 38% receiving ESA funds.
• Why they buy: “I missed three inquiries last weekend. Two of them visited [competitor school] before I called Monday. We’re 8 students under our enrollment target — that’s $120K of tuition.”
• TAM reasoning: 87% of US private schools (~33K) have <300 students → ~29K schools. Plus ~95K microschools, of which the ~38% on ESAs (~36K) are most equipped to pay. Realistic serviceable target: 60K orgs.
• Why now for them: ESA money is hitting wallets — parents who couldn’t afford private school 18 months ago are now shopping. Schools that used to get 30 inquiries/yr are getting 100+. Most still have one part-time admissions person. The bottleneck is the front of the funnel, not the close.

5. Product sketch (MVP)

• 60-second AI responder, every channel — answers web-form, email, SMS, FB Messenger, and inbound phone with a school-voice reply, day or night
• Auto-book a tour — voice/chat agent picks an open slot from the head-of-school’s Google/Outlook calendar, sends ICS + confirmation
• Pre-tour family briefing — auto-drafts a personalized “what to know about [School] before your visit Tuesday” email/SMS the night before, tied to the family’s stated reasons for inquiring
• Drop-off radar — flags any family that’s gone quiet >5 days post-inquiry or post-tour and drafts a personalized nudge for the director to approve in one click
• Monday digest — a Mon-7am email to the head of school: hot families, cold families, this week’s tours, each with a one-line action
• Light onboarding — paste your About page, mission, fee schedule, FAQ, schedule of tours; we tune the agent’s voice in 30 minutes
• Bolts on, doesn’t replace — pushes contact records to SchoolCues / Blackbaud / Finalsite / Google Sheets via Zapier or simple webhook; schools keep their existing SIS

Explicitly NOT v1: gradebook, billing, attendance, application portal, financial aid scoring, full student profiles. Stay narrow.

6. AI angle — what’s load-bearing

Three places the AI is doing actual work, not decorating:

1. Voice agent that handles a parent saying “Hi, I have a 7-year-old, we’re moving from Austin, do you have spots in 2nd grade for Fall, and how does your school handle dyslexia?” — and books a tour while answering accurately, in the school’s voice, without hallucinating the dyslexia program.
2. Personalization at scale. A real admissions director writes a different follow-up email to a Montessori-curious-secular family than to a homeschool family asking about classical Latin. The LLM does that for 100 families/week without sounding like spam.
3. Drop-off radar. Reads the thread + tour notes, decides “this family went cold because they’re price-shopping” vs. “this family went cold because nobody addressed their dyslexia question” — and drafts the right re-engagement.

Without AI, this is a $40K/yr human admissions counselor. AI is the entire reason a school under 300 students can afford this.

7. Localization angle (if any)

N/A — this is a US-first play. School-choice ESA programs are a uniquely US regulatory tailwind, and small private K-12 distribution channels (ACSI, NCEA, NAIS regional, classical school networks) are US-specific. Worth revisiting for UK, Canada, and Australia indie/Catholic markets in year 2.

8. Business model — path to $1M–$5M ARR

• Pricing tiers:
  • Microschool / under 100 students: $149/mo, single campus, 1,500 voice-min, unlimited chat/email
  • Small school (100–300): $299/mo, 3,000 voice-min, drop-off radar, multi-user
  • Growth (300–600): $599/mo, 8,000 voice-min, multi-campus, integration to Blackbaud/Finalsite/Veracross
  • Diocese / network / classical group (5+ campuses): $1,500–$4,000/mo via annual contract
• ACV math: Average ~$3,500. Microschools at $1,800, small schools $3,600, growth $7,200, networks $25K+.
• $1M ARR: ~280 customers at blended $3,500 ACV. Captures <0.5% of the 60K-org SOM.
• $5M ARR: ~1,400 customers — still <2.5% market penetration. Realistic in 30–36 months given ESA tailwind.
• Expansion path: add application-portal autofill ($50/mo), tuition-aid pre-screening, donor/alumni re-engagement (same agent, different mission). Move into private daycare/preschool follow-up (same playbook, different sales motion). Cross-sell to networks of 10–80 schools (Catholic dioceses, ACCS, classical co-ops).

9. Go-to-market wedge — first 100 customers

Concrete plays, not “we’ll do SEO”:

1. ACCS / SCL / classical-school conferences (June + Oct 2026). Two-thirds of classical Christian schools are ACCS or SCL members, all are <300 students, all desperately under-resourced on admissions, all have a tight grapevine. Booth + heads-of-school dinner at one conference = 30+ qualified pipeline. Tier-1 channel.
2. Diocese-level deals. Texas Catholic Conference + Florida Catholic Conference are organizing for the 2026-27 ESA enrollment surge. One sale = 8–60 schools. Even 2 dioceses = 100+ campuses.
3. Private-school admissions director Facebook groups. “Independent School Admissions Officers”, “ACSI Admissions Pros”, “Catholic School Admissions” — 5K–15K-member private groups, surprisingly active. Target a 7-day “speed-to-lead audit” lead magnet (we audit your response time + show how many leads you lost). Convert to trial.
4. Cold outreach to the 5,000 schools that actually have an inquiry form on their website. Scrape, send a 90-second Loom showing the AI booking a tour from their actual form vs. their actual current 36-hour response. 1% close = 50 customers. Tested by similar AI-voice-for-vertical playbooks.
5. Microschool-network partnerships. Prenda (1,000+ founders), KaiPod (150+), Acton (300+), Wildflower (500+). Revenue-share with the network operator who already has the trust.

10. Build complexity — justification

Medium. Off-the-shelf voice (Retell or Vapi for the agent runtime, Deepgram/ElevenLabs/Cartesia for STT/TTS), Twilio for telephony + SMS, OpenAI/Anthropic for reasoning, Google Calendar / Microsoft Graph for booking, Postmark for email, off-the-shelf web chat embed. Custom work: school-onboarding flow (turn About-page + handbook into agent persona), school-specific guardrails so the voice agent doesn’t hallucinate programs that don’t exist, FERPA-aware data handling, drop-off-radar logic, and the Monday digest. Pair could ship v1 in 12–14 weeks; integration polish + first FERPA review pushes to 16–18.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	FERPA-relevant; we don’t store student records, just inquiries. State-by-state telephony recording disclosure handled.
Ethical — no harm / dark patterns	✅	Agent identifies as AI; no deceptive enrollment tactics.
Market exists (evidence above)	✅	29K small private + 95K microschools; ESA tailwind; quantified speed-to-lead loss.
1–5 person team can build this	✅	Pair of generalists in 12–16 weeks.
Launchable with <$50K / ₹40L	✅	Mostly variable LLM/voice cost; ~$15–25K to first paying cohort.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16	Real money on the line — 1 lost family = $14K–$130K LTV. Felt acutely during AEP-equivalent (Jan–Apr inquiry season). Director knows leads are leaking but can’t fix it alone.
Demand evidence	15	12	Quantified speed-to-lead stats; 70% drop-off; 77% of schools using only free tools; ESA boom. Direct customer-voice quotes from director forums. Generic voice-agent vendors emerging but not vertical-tuned.
Build feasibility	15	11	Voice + LLM + calendar + email orchestration is well-trodden. School-specific guardrails and FERPA review are the gnarly bits. Medium not Low.
Distribution clarity	15	11	Concrete: ACCS/SCL conferences, dioceses, FB groups, scrape-and-Loom outreach, microschool-network partnerships. Cycle is school-year-paced; need to time launch for Mar–May (next-year planning).
Revenue mechanics	15	12	Pricing benchmarked to category ($150–$600/mo) and ROI is one saved enrollment. ACV $3.5K. $1M ARR at ~280 customers is plausible inside 12 months given ESA wave.
Time to first revenue	10	8	6–10 weeks: book pilots from FB-group leads pre-launch, paid pilots at $99/mo first 3 months, convert to standard pricing.
Defensibility	10	8	Soft moats: school-voice tuning becomes data flywheel, integrations into ACSI/NCEA distribution lists, “speed-to-lead audit” content owns the SEO term. Easy to copy technically; hard to copy distribution + niche credibility.
Total	100	78

13. Qualitative modifiers

Founder-fit tags

technical-heavy · sales-heavy

This is a builder + a school-savvy seller. The product is doable, but it lives or dies on whether the seller can get into ACSI/NCEA conference circuits and run the trust-based motion. A solo technical founder will struggle. Best fit: ex-edtech operator + AI-voice engineer.

Key assumptions to validate (3–5)

1. Assumption: Small school admissions directors will pay $299/mo for a focused tool that bolts on top of their existing SIS, instead of waiting for SchoolCues/Finalsite to add AI. How to test: 30 cold-outbound demos to admissions directors at ACSI member schools in the next 4 weeks. Target: 5 paid pilots at $149/mo.
2. Assumption: Voice-AI quality in 2026 is good enough for a parent-facing first impression of a Christian/classical school without embarrassing hallucinations. How to test: Build a 1-week prototype, have 10 admissions directors live-call it about their school after a 30-min onboarding; measure “would you put this on your website?” yes/no.
3. Assumption: The Monday-digest + drop-off-radar is the sticky feature, not the voice agent. How to test: Track week-4 retention among first 20 pilots — which feature do they self-report as “I’d cancel if you removed it.”
4. Assumption: Diocese-level deals (one sale = 10–60 schools) are accessible to a tiny new vendor. How to test: 5 in-person meetings with Catholic-conference superintendents in TX, FL, IN by Sep 2026. If 0 of 5 advance to pilot, the diocese motion isn’t real and we re-bias to direct.

Risk flags

1. Platform/AI dependency: voice quality, latency, and pricing depend on Retell/Vapi/OpenAI/Anthropic. A 2× price hike or a regression destroys margins. Mitigation: build the abstraction layer + multi-provider routing from day 1.
2. Sales-cycle drag at network level: dioceses and ACCS networks move on academic-year calendars. Don’t underwrite a 90-day-cycle business on diocese revenue; bottoms-up direct must work first.
3. FERPA / state telephony recording: states are split on one-party vs. two-party consent for voice agents. Need a clean disclosure flow and per-state config. Underestimating this can kill a pilot in California or Florida.
4. Incumbent “AI feature” launch: SchoolCues, Finalsite, Blackbaud will eventually bolt on a voice agent. Speed and vertical depth matter — we have to be the obvious “AI for small schools” answer before they catch up.

14. Structured verdict

Score:                  78/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       AI-voice engineer + ex-edtech / private-school operator
Time to revenue:        6–10 weeks (paid pilots), $1M ARR plausible in 14–18 months
Capital to launch:      $15–25K to first paying cohort; $40–50K to 50 customers
Top 3 assumptions to validate first:
  1. 30 cold-outbound demos → 5 paid pilots at $149/mo within 4 weeks
  2. Live-call test with 10 admissions directors: >7 of 10 say "yes, I'd put this on my site"
  3. Diocese motion: 5 superintendent meetings, ≥1 advances to pilot in 90 days
Kill criteria:
  - Abandon if <3 of 30 cold demos convert to paid pilot
  - Abandon if voice-agent live-call fails the "would you put this on your site" test (≤4 of 10 say yes)
  - Abandon if SchoolCues + Finalsite both ship native AI voice for under $99/mo before our v1

15. Next step — 1-week validation sprint

• Day 1–2: Build a stripped-down Retell + GPT-5 + Google Calendar agent loaded with the actual website + handbook of one volunteer ACSI school. No code shipped publicly.
• Day 3: Have the volunteer admissions director call it 5 times pretending to be different types of parents (price-shopper, dyslexia-curious, classical-curious, secular-curious, transfer-from-public). Record outputs. Measure: would the school put this on their site as-is?
• Day 4: Run 15 cold-outreach 90-sec Looms to admissions directors at ACSI/NCEA schools showing “we just answered this inquiry on your website in 60 seconds — here’s the recording.”
• Day 5: Decide go / no-go. Falsifiable bar: ≥3 of 15 schools book a paid-pilot intro call AND the volunteer director rates the live-call test ≥7/10 on “would put on site.” If either fails, re-bias to a different vertical (microschool networks first, or Catholic-only) before committing.