SiteLine — AI voice receptionist for independent US RV parks

1. One-liner

AI phone agent that answers every RV-park call, checks live availability, and books the site mid-call.

2. Trend signal — why now?

Three forces converged in the last 12 months.

(a) Voice-agent unit economics collapsed. Retell AI publishes a flat $0.07/min voice agent rate. Synthflow lists $0.08/min bundled. Vapi headline $0.05/min (with stacked fees). A 4-minute campground call now costs $0.28–$0.40 in inference — under a tenth of a human answering service.

(b) Mom-and-pop campgrounds are getting crushed by missed calls. ALM Corp’s 2026 SMB study cites 62% of calls to small businesses go unanswered, 85% of those callers never try again, and the average SMB loses ~$126K/yr to missed calls. RV parks are the textbook case: the owner is mowing site 14, the office phone rings, voicemail picks up, the caller is in a Class A on I-40 looking for tonight — they call the next pin on the map.

(c) PE consolidation of campground PMS is pushing owners to bolt-ons. Storable owns Newbook. Campspot is venture-backed (per PitchBook profile). 127 parks were rolled up by PE in 2023 alone; the Blue Metric $97M seven-park deal closed in 2025. Independent operators (78% of the ~15,000 US parks) are watching incumbent PMS prices creep and looking for cheaper plug-ins, not a $7,500-setup-fee API rebuild.

Forum sentiment matches:

• “by the time we get around to some of them the people have made their reservations elsewhere” — campground owner, RV LIFE forum, on manual reservation entry from website forms
• “performs like software did in 1997 (very cumbersome and not intuitive)” — captainehh, MobileHomeUniversity forum, on a legacy PMS
• “incredibly un-user friendly and lacking in certain features” — Suburban, MobileHomeUniversity forum, on Campspot

Provenance:

• Signal 1 (demand): 62% of SMB calls unanswered, $126K/yr lost per SMB — https://almcorp.com/blog/best-ai-receptionist-products-2026/ — 2026-05-13
• Signal 2 (feasibility): Retell $0.07/min, Synthflow $0.08/min, Vapi $0.05/min flat — https://www.retellai.com/blog/ai-voice-agent-pricing-full-cost-breakdown-platform-comparison-roi-analysis — 2026-05-13
• Signal 3 (economic): PE rolling up campgrounds, Blue Metric $97M deal, Storable owns Newbook — https://ohweekly.substack.com/p/blue-metric-groups-97m-rv-portfolio + https://www.newbook.cloud/our-platform/channel-manager/ — 2026-05-13 Category: Underserved niche × Tech-unlock

3. The opportunity

Campground PMS players (Campspot, Newbook, RoverPass, CampLife, Firefly, Bonfire) are racing to add AI receptionist features as part of a full-PMS replacement. That is the wrong shape for the 12,000 independent parks already locked into a PMS, paper books, or Google Calendar. None of them want a rip-and-replace migration in shoulder season.

The wedge is bolt-on voice, not bolt-on PMS. SiteLine plugs into whatever the park already runs — Campspot, Newbook, RoverPass, Firefly, even Google Sheets / iCal / paper-and-confirm-back — and answers the phone with a campground-fluent agent that:

• knows the difference between a 30-amp back-in and a 50-amp big-rig pull-through,
• can quote this park’s rates including weekly/monthly/seasonal,
• takes the deposit via SMS link mid-call (Stripe / Square),
• writes the reservation back to the PMS or holds it for owner confirmation if no API.

Generic AI receptionists (Rosie $49+, Goodcall $79+, RingConnect, NextPhone, Allo) are field-services-flavored. They will book a plumber. They do not know what “sani-dump” means. A vertical-tuned product wins on every metric the owner cares about — call resolution rate, mis-quote rate, abandonment rate.

4. Target market

• Primary customer: Owner-operator of an independent US RV park or campground, 30–200 sites, 1–3 properties. Often a retiree couple or a second-career operator. Already on a PMS but bleeding calls, OR off-PMS and bleeding both.
• Why they buy: “The phone rings while I’m mowing. By the time I’m back to the office, they’ve called Jellystone.” Real, daily, measurable lost revenue. Average campground nightly rate ~$45–$80. A single saved call/week = $2K+/year in incremental revenue, well past the SaaS line.
• Rough TAM reasoning: ~15,000 US RV parks + campgrounds. ~78% independent = ~11,700. Strip out the bottom quartile (too small / paper-only and never coming online) and the corporate top — addressable ~8,000–9,000 parks. Adjacent: glamping (~3,000 in US), youth/church camps (~5,000), state-park concessionaires.
• Why now for them: Camping demand softened in 2025 (Hipcamp: “uneven, families more careful with budgets”). Owners feel every lost site. Meanwhile their PMS price is creeping post-PE-acquisition. They want a cheap, high-ROI lever they can flip this season.

5. Product sketch (MVP)

• Local US phone number (or port the park’s existing 800/local) that AI answers 24/7 in seconds
• Campground-fluent voice persona: full-hookup, back-in, pull-through, big-rig, 30/50-amp, sani-dump, holding tank, weekly/monthly rate
• Live availability check against Campspot / Newbook / RoverPass API where available, iCal/Google Calendar fallback elsewhere, “human will confirm by SMS” fallback for paper books
• Mid-call SMS deposit link (Stripe/Square), confirmation back to caller before hang-up
• Park-specific FAQ memory: pet policy, age of allowed rigs, generator quiet hours, propane availability, nearest dump station — owner trains it once in a 20-minute setup
• Email + SMS digest to owner each morning: “8 calls answered, 5 booked ($1,840), 2 needed callback, 1 directed to voicemail” with transcripts
• Park-branded greeting and on-hold voice in the owner’s actual recorded voice (ElevenLabs clone, ~3 min sample)
• Spanish handoff for ESL callers (huge in TX/CA/FL/AZ snowbird traffic)

6. AI angle — what’s load-bearing

Without AI this product is impossible. The whole point is a sub-1.5s-latency, multi-turn voice agent that:

1. Transcribes campground-noisy calls (callers on cell in moving rigs, road wind, kids in background) — Whisper-class STT.
2. Reasons over campground vocabulary, the park’s specific rate sheet, this week’s site availability, and policy edge cases (“can we bring our 45-foot toy hauler?”) — LLM with retrieval over the park’s loaded config.
3. Drives a structured booking action mid-call — function-calling into the PMS API or the iCal fallback.

Strip the AI and you have a phone tree — i.e. exactly what the owner has now, the one that loses 62% of calls. The AI is the product.

7. Localization angle (if any)

US-first, English + Spanish. The Spanish wedge alone is meaningful — south-Texas, Florida, Arizona, and California snowbird traffic includes a non-trivial Hispanic share, and zero current campground PMS voice systems handle Spanish well. After v1, Canadian English + French Canadian opens ~3,000 Canadian parks. UK caravan/holiday parks, AU/NZ holiday parks are eventual adjacencies — different vocabulary (“touring pitch”, “powered site”), same architecture.

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

• Pricing tiers:
  • Starter $99/mo — up to 200 answered calls/mo, 1 park, iCal/manual confirm only
  • Pro $199/mo — up to 600 answered calls/mo, PMS integration (Campspot/Newbook/RoverPass/Firefly), Spanish, owner-voice clone
  • Multi-park $349–$549/mo — 2–3 properties, central dashboard, fleet of numbers
  • Overage: $0.20 per call beyond plan cap (still 50%+ gross margin at Retell’s $0.07/min)
• ACV: Blended ~$1,800/yr (most parks land on Pro)
• Path to $1M ARR: 555 parks × $149/mo blended × 12 = $1.0M. ~5% of addressable ~11,700 independents.
• Path to $5M ARR: ~2,300 parks + multi-park upgrades + glamping/youth-camp adjacency. Requires moving past direct-to-owner into channel partnerships (Newbook resell, state campground associations).
• Expansion path: SMS marketing reactivation (texting past callers who didn’t book), upsell premium voices, post-stay review-collection calls, lost-reservation reactivation, eventually outbound winter snowbird booking calls.

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

1. OHCE 2026 in Fort Worth, Nov 9–11 — $129 expo pass. ~3,000 attendees, almost all of them buyer-owners. A 10x10 booth with a live demo phone (“call this number, try to break it”) closes 30–50 if the product works. Pre-conference: cold-email the published exhibitor list.
2. State campground associations (TXCOA, FloridaARVC, CalARVC, GeorgiaCOA, etc.) — 35+ in the US. Pay $1K–$3K to sponsor a webinar to their member list. Each list is 100–800 owners. Conversion math: 1% paid trial = 1–8 customers per webinar. Run 8 in the first 6 months.
3. Modern Campground + Woodall’s Campground Magazine — both publish weekly e-newsletters to ~25K outdoor-hospitality decision-makers. Sponsor + content collaboration ($1K–$2K/issue).
4. Cold outbound to Campspot/Newbook complaint posters — scrape Capterra, G2, Software Advice reviews of the major PMS players (Campspot 200+ reviews, Newbook 100+). DM/email the named reviewers who complained about service/pricing, offer 30-day free trial.
5. PMS partnership — Firefly, Bonfire, CampLife are too small to build their own voice product. Co-sell deal: SiteLine pays 20% revshare for referrals. Aligns incentives — PMS doesn’t lose seats, gains a sticky bolt-on.

Demo strategy: ship a public “call this number to try the demo” with a fake “Pine Ridge Campground” persona. Lets prospects feel it before any sales conversation.

10. Build complexity — justification

Medium. Off-the-shelf stack: Retell or Vapi for voice (won’t build the speech layer), Twilio for telephony, Stripe/Square for deposits, OpenAI/Claude for reasoning, ElevenLabs for owner-voice clones. Custom work: (a) campground vocabulary/rate-sheet ingestion + RAG, (b) per-PMS connectors (Campspot, Newbook, RoverPass, Firefly — each ~1–2 weeks), (c) owner dashboard + transcript review, (d) iCal/Google Cal fallback path. Pair of builders, v1 in ~12 weeks, polished GA in ~16. The connectors are where the calendar blows — accept it.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Standard AI voice disclosure (“calls may be recorded for service”) + state TCPA-equivalents handled.
Ethical — no harm / dark patterns	✅	Agent identifies as AI on first turn; opt-out to human callback always offered.
Market exists (evidence above)	✅	$126K/yr/SMB lost-call cost + 11,700 independent US parks + forum quotes confirm.
1–5 person team can build this	✅	2 engineers, 12–16 weeks to GA on off-the-shelf voice stack.
Launchable with <$50K / ₹40L	✅	~$15K for v1: voice agent credits, Twilio numbers, a few PMS sandbox subs, basic landing site.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	18/20	Daily, measurable revenue loss. Owners can quote the exact number of calls they missed last weekend.
Demand evidence	15	13/15	Multiple PMS competitors, generic AI receptionists with paid traction, public lost-call stats, forum complaints. Missing one strong signal: a direct “I’d pay for this” quote from a campground owner specifically. Validate in week 1.
Build feasibility	15	12/15	Off-the-shelf voice stack handles 80%. The 20% — PMS connectors and per-park config — is real engineering.
Distribution clarity	15	13/15	OHCE + state associations + Modern Campground media + cold outbound to PMS complainers is a named, costed playbook.
Revenue mechanics	15	12/15	Pricing benchmarked vs Rosie/Goodcall + below human answering services. ACV/customer-count math credible. Risk: usage overage caps eat margin if a viral park gets 2K calls/mo.
Time to first revenue	10	8/10	6–8 weeks to first paying park if you pre-sell at a state association webinar pre-launch. Without that, 10–12 weeks.
Defensibility	10	7/10	Workflow lock-in (owner trains voice + FAQ once, painful to redo), PMS integration depth, campground-vocabulary brand. Not a fortress; a focused 6–12 month head start.
Total	100	80/100	STRONG GO.

13. Qualitative modifiers

Founder-fit tags

technical-heavy (voice agents + PMS connectors), domain-expertise-required (must speak campground; ideally a co-founder who has owned or worked at a park, or a 2-month immersion). A non-domain founder can still ship this but must front-load 30+ owner interviews before line one of code.

Key assumptions to validate (3–5)

1. Assumption: Owners will pay $149–$199/mo for a bolt-on (vs replacing PMS). How to test: 30 owner phone interviews, ask the exact price with a credit-card-pre-auth ask at end. Looking for ≥10 willing to put $1 hold to reserve.
2. Assumption: Voice agent quotes correct rates and reads PMS availability accurately ≥95% of the time on real campground calls. How to test: Wizard-of-Oz at 3 friendly pilot parks for 2 weeks; transcript-audit every call.
3. Assumption: Campspot/Newbook/RoverPass/Firefly will let a bolt-on read availability without an enterprise contract (Campspot’s published API is $7.5K setup — pricing here matters). How to test: Direct outreach to each PMS BD team. Identify minimum 2 viable PMS paths + iCal fallback before commit.
4. Assumption: Owner-trust hurdle is crossable — i.e. owners are willing to let “AI book a campsite” without their human review. How to test: A/B “auto-book” vs “AI proposes, SMS owner to confirm” in the first 10 pilots. If only the SMS-confirm path lands, that’s still a viable product, just a thinner one.

Risk flags

1. Platform dependency on voice agent vendor. Retell/Vapi/Synthflow could change pricing or terms. Mitigation: design the orchestration layer to be voice-vendor-agnostic, run pilot on Retell + parallel test on Vapi.
2. Storable / Campspot launches a competing in-PMS AI receptionist. Likely within 12–18 months. Mitigation: (a) get to 200+ parks fast on bolt-on speed, (b) be cross-PMS, which neither incumbent will ever be, (c) explore reseller deal with Firefly/Bonfire/CampLife (smaller PMS players who can’t build their own).
3. Seasonality risk on usage. Northern parks are dead Nov–Mar. Plans must accommodate seasonal pause/resume, or churn will be brutal in winter.
4. TCPA / call-recording compliance. Multi-state recording rules differ. Mitigation: default to two-party-consent disclosure in greeting.

14. Structured verdict

Score:                  80/100
Verdict:                STRONG GO
Confidence:             High
Best-fit builder:       Technical co-founder + a co-founder/advisor who owned or
                        worked at an RV park; alternative is a solo technical
                        founder who lives in the parks for 60 days pre-build
Time to revenue:        6–12 weeks (faster if pre-sold at a state-association webinar)
Capital to launch:      ₹12L / $15K
Top 3 assumptions to validate first:
  1. Owner willingness-to-pay $149–$199/mo (30 interviews + soft pre-auth)
  2. Voice agent rate/availability accuracy ≥95% on real calls (3-park WoZ pilot)
  3. PMS read-access viability (BD calls with Campspot/Newbook/RoverPass/Firefly)
Kill criteria:
  - Abandon if <8 of 30 owner interviews show $149+/mo willingness
  - Abandon if voice accuracy <90% after 2 weeks of pilot tuning
  - Abandon if Storable ships a free in-Newbook AI receptionist before MVP
  - Abandon if no PMS partner agrees to read-access at <$2K/year by month 3

15. Next step — 1-week validation sprint

• Day 1–2: Build a one-page landing site with a live “call this number” demo persona (“Pine Ridge Campground”). Wire to a basic Retell agent loaded with a fake rate sheet + 30-site fake availability. Cost: $200.
• Day 3–4: Cold-email 200 RV park owners scraped from OHI member directory + Capterra Campspot/Newbook reviewers. Pitch the live demo line. Target 25 to actually call the demo, 15 to take a 20-minute interview.
• Day 5: Run 12+ interviews. Two questions matter: (a) “If this worked for your park, would you pay $199/mo?” (b) “Can I put a $1 pre-auth on your card to hold a launch spot?”
• Decision: Go if ≥8/12 say yes to (a) AND ≥4/12 give the $1 pre-auth. Anything less = back to Stage 2.

The falsifiable signal is the $1 hold. Talk is cheap. A card on file is not.