Alonvia Pilot — Decisions Needed from Jim
Overview
From: Josh · Date: 2026-08-07 · Companion to: Marketing-Engine-PRD.md
Three decisions are blocking the pitch and the build. Everything else can wait. Each section ends with exactly what I need from you.
Decision 1 — Price point, and what it honestly buys
You want the lowest price that still delivers value, for a skeptical room. Agreed. But before you pick a number, here is the real cost of running this, because the price determines what we can promise without over-promising — and whatever we promise to ten friends from the study group, we will actually have to do.
The effort math, straight
Per-client onboarding (first 30 days) — roughly 10–15 hours each, even with our tooling: audit + review, chasing the practice for logins/photos/GBP access/ad-account setup (this chase is reliably the single biggest time sink with dental offices), landing page build + your review, campaign build + approvals, technical GEO fixes. Ten clients signing around the same time = 100–150 hours of onboarding in the first month or two, on top of Josh building the tooling itself.
Per-client ongoing — 2–4 hours/month at maturity, more early: daily monitoring approvals, monthly optimization pass, monthly report, and client comms. This last one deserves emphasis: these are warm-network clients. They will text you. "Quick questions" from friends are a real recurring cost, and they don't show up in any feature list.
Multipliers we've already committed to: multi-state means ten separate local markets — research and compliance checks don't amortize across clients. Bilingual-as-P0 means roughly 1.5–2× the content effort on every page and ad.
What each price honestly delivers
| $299/mo | $499/mo | $299/mo + $750 one-time setup | |
|---|---|---|---|
| Covers | A hard-fenced minimum: audit + baseline, technical GEO fixes, one landing page (EN + ZH), one Google campaign, monthly report. Monthly-cadence touch, email-only support. | Everything at left, plus a real monthly optimization cycle and ~2 GEO content pages/month. | Same as $299 tier, but the setup fee covers the actual front-loaded onboarding cost. |
| The truth about it | A loss-leader. $3k MRR across 10 clients does not pay for the hours above. Fine if we treat it as buying proof and testimonials — but then the scope fence must be in writing, because saying no to extras is hardest with friends. | Closer to sustainable, still cheap against the $1.5k–5k/mo agencies charge. | The most honest structure, because our effort is front-loaded. Question is whether a setup fee kills the "easy yes" for this room. A 3-month prepay (~$900) is a softer equivalent. |
Not deliverable at any of these prices, and must not be implied in the room: website rebuilds, unlimited revisions, phone-support SLA, review management, Chinese-platform management (Xiaohongshu/WeChat), guaranteed lead counts, or any promise that AI engines will recommend the practice. These are future tiers or not promises at all.
Need from you: pick a column (or propose a number), and sign off on the excluded list above going into the service agreement verbatim. My lean: $299 + setup fee or 3-month prepay — but any column works if the fence holds.
Decision 2 — First ad platform
Recommendation: Google only (Search + Google Business Profile) for the pilot. Implant patients search with intent — "implant dentist near me" — and that intent lives on Google. Meta is interruption advertising, weaker for high-intent high-ticket procedures, and every added platform multiplies creative formats, monitoring load, and integration work against the same low price point. Meta and Chinese platforms become future-tier upsells.
Need from you: sign-off on Google-only, or tell me what you'd add and we re-run the Decision 1 math, because it changes.
Decision 3 — AI in the patient-facing and scheduling layer
You want an agent answering leads and booking appointments, to prove our efficiency. Here is the unbiased version.
The case for it
- Speed-to-lead is the single biggest conversion lever in dental marketing. Leads contacted within minutes convert at multiples of leads contacted hours later, and dental front desks miss a large share of calls — especially after hours, which is when people research implants. An agent that responds in 30 seconds, at 9pm, in English or Chinese, is real value, not a gimmick. - It protects our results. If our campaigns generate leads that a busy front desk never calls back, our numbers look bad through no fault of the campaigns. Owning first contact closes the attribution loop — we can show "we delivered N responsive leads," not "we spent your budget." - It's the efficiency demo. Nothing in the deck proves the AI-team thesis like a dentist watching a lead get answered instantly. It's also the natural future upsell tier, and it's sticky.
The case against it
- It crosses the PHI line the pilot deliberately stays behind. The moment a prospective patient tells our agent about their teeth, we're handling health information: BAAs with every practice, HIPAA-eligible AI infrastructure, retention and encryption policies, legal exposure. Not impossible — but it converts a marketing pilot into a healthcare-software pilot. - Live conversation breaks our own safety model. Our rule is that clinical content is human-written. A conversational agent will get asked clinical questions ("can I get implants with my diabetes?") and cannot be pre-approved message-by-message. Guardrails reduce but don't eliminate failure, and one bad screenshot travels through the entire 500-dentist network we're counting on. There's also the tail risk that matters most: a patient messaging something urgent ("my implant site won't stop bleeding") and a bot mishandling it. - Scheduling integration is a known trap. Writing into practice systems (Dentrix, Eaglesoft, Open Dental) is months of integration work plus BAAs — the business plan literally flagged this pattern as "reaching for infrastructure where a phone call suffices." Even without PMS work, it needs buy-in from front-desk staff, who tend to resent and quietly undermine tools that touch their book. - The effort math from Decision 1 doesn't survive it. This is a second product. At $299–499/mo it's dramatically underpriced, and building it now consumes the exact dev and monitoring capacity the core loop needs. Our pilot metric is revenue and operator leverage, not feature count.
The middle path, and my recommendation
There's a thin version that captures most of the conversion value with little of the risk:
- Instant structured lead response — missed-call-text-back and form follow-up that collects name, phone, preferred time, preferred language, and says the office will confirm. No health questions asked, none answered — clinical questions get a canned "great question for the doctor" deflection. Stays essentially out of PHI.
- A booking link on our landing pages (Calendly-style consult slots the practice confirms) — scheduling without PMS integration.
Recommendation: include the thin version in the pilot; defer the full conversational agent and PMS-integrated scheduling to a post-pilot paid tier, gated on BAAs and HIPAA-eligible infrastructure. The thin version still demos "a lead got answered in 30 seconds at 9pm" — the efficiency proof you want — while the full version, built now, is the single likeliest thing to blow the timeline, the price point, and possibly a friendship's worth of trust if it misfires. If the pilot converts, the full agent becomes the flagship of tier 2, sold to clients who already trust us.
Need from you: yes/no on the thin version in pilot scope, and agreement (or pushback) on deferring the full agent. If you want the full agent anyway, I need us to re-scope timeline and price together first — it roughly doubles the build.
Answers get folded into the PRD (§4, §5, §11) and the service agreement. — J