The flow is simple enough to build in an hour and robust enough to run while you travel.

End‑to‑end path:

1. Entry: Click‑to‑WhatsApp ad, QR code, or wa.me link →
2. Greeting + permission/opt‑in →
3. 4‑question qualifier (Budget, Timeline, Use case, Authority) →
4. Offer route: Book call (Calendly/TidyCal) OR pay deposit (Stripe Payment Link) OR workshop invite OR nurture →
5. Confirmations + reminders →
6. CRM/log update →
7. Human handoff on edge cases/errors.

Nomad defaults: keep responses templated, let Calendly own time zones, and set monitoring to catch issues while you’re on planes.

Pick any BSP you prefer; the logic below is provider‑agnostic.

Prereqs:

• WhatsApp Business Platform access + dedicated number
• BSP account (e.g., Gupshup/Twilio/WATI) with template submission enabled
• Automation runner (Make.com or n8n)
• Calendar app (Calendly/TidyCal) with API access
• Payment processor (Stripe Payment Links)
• LLM API (e.g., GPT‑4o‑mini or equivalent) for light intent parsing
• Alert channel (Telegram/Email) for handoff

Blueprint (Make.com or n8n modules):

1. Trigger: Incoming WhatsApp message webhook → store [leadid, wanumber, source, timestamp, thread_id].
2. Consent Gate: If no explicit permission recorded, send [OPTINREQUEST] template; proceed only on affirmative intent.
3. State Machine: Manage conversation state (Q1→Q4→Route). Enforce allowed replies (A/B/C) to keep the LLM minimal.
4. Questions: Send templated messages Q1–Q4; capture normalized answers.
5. Scoring + Routing: Evaluate Budget/Timeline/Use‑case/Authority → set route = [BOOKFREECALL | BOOKPAIDCALL | WORKSHOP | NURTURE | HUMAN].
6. Calendar: Fetch available slots via Calendly API (owner = you). Create prefilled scheduling link with [name, email/WA, timezone auto].
7. Payments (optional): Generate Stripe Payment Link for [DEPOSITAMOUNT] when route = BOOKPAID_CALL.
8. Messaging: Deliver booking/payment link template; on success, send confirmation + reminders (24h and 2h before).
9. CRM/Sheet: Upsert lead with answers, route, booking URL, payment status; append transcript pointer.
10. Monitoring: Daily digest of counts/costs; real‑time error to Telegram if state stalls >10 min, payment fails, or non‑matched intent → HUMAN_HANDOFF.

Copy these as WhatsApp templates (keep variables in [BRACKETS]). Submit 2–3 variants per critical step to survive rejections.

1. OPTINREQUEST

“Hey [FIRST_NAME] — you tapped my WhatsApp link from [SOURCE]. I help [ICP] with [OFFER]. OK to ask 4 quick questions to see if I can help? Reply YES to continue or STOP to opt out.”

1. Q1_BUDGET

“To recommend the right path, where’s monthly revenue? A) < $5k B) $5k–$15k C) $15k+ (Reply A, B, or C.)”

1. Q2_TIMELINE

“When do you want this live? A) This month B) Next quarter C) Just researching (Reply A, B, or C.)”

1. Q3_USECASE

“What’s the main thing you want to automate? A) Lead capture B) Client onboarding C) Content repurposing D) Customer support (Reply A–D.)”

1. Q4_AUTHORITY

“Are you the decision maker for this? A) Yes B) I share decision C) No (Reply A, B, or C.)”

1. ROUTEBOOKFREE

“Looks like we can help. Grab a time here: [CALENDLY_LINK]. It shows your local timezone. You’ll get reminders automatically.”

1. ROUTEBOOKPAID

“We can fast‑track you with a credited strategy session. Secure your spot with a [DEPOSITAMOUNT] deposit here: [STRIPELINK]. It’s fully credited toward any package you choose.”

1. ROUTE_WORKSHOP

“Best next step: our [WORKSHOPNAME] this week for $[PRICE]. Seat link: [CHECKOUTLINK]. You’ll leave with a working mini‑system.”

1. CONFIRM_BOOKING

“Locked in: [DATELOCAL] at [TIMELOCAL]. You’ll get a 24h and 2h reminder here. Need to reschedule? Use this link: [RESCHEDULE_LINK].”

1. REMINDER_24H

“Reminder: Tomorrow [DATELOCAL] at [TIMELOCAL]. Bring: [PREPITEMS]. Reschedule: [RESCHEDULELINK].”

1. EDGECASE_HANDOFF

“That’s a great question. Looping in [HUMANNAME]. Expect a reply within 12 hours. Want to book instead? [CALENDLYLINK]”

1. OPT_OUT

“No problem — you’re unsubscribed. If that changes, you can message START any time.”

Prompt guardrail (system prompt for LLM):
“You are a WhatsApp concierge for [BRAND]. You only send pre‑approved templates with variables filled. Never invent times, prices, or claims. If a user’s message doesn’t map to an expected answer (A/B/C/D or YES/STOP), respond with EDGECASE_HANDOFF and raise an alert.”

Translate answers to actions with explicit rules so the bot can’t wander. Recommended defaults for solo nomads:

Budget (Q1): A (<$5k) → WORKSHOP; B ($5k–$15k) → BOOKFREECALL; C ($15k+) → BOOKPAIDCALL with [DEPOSIT_AMOUNT=$97].

Timeline (Q2): If C (Researching) → NURTURE regardless of budget; else continue.

Use case (Q3): Tag intent for agenda + case studies in reminders.

Authority (Q4): If B or C (shared/no), append “Include decision maker?” step; else continue.

Routing pseudo‑logic:
```
if TIMELINE == RESEARCH: route = NURTURE
else if BUDGET == C: route = BOOKPAIDCALL
else if BUDGET == B: route = BOOKFREECALL
else: route = WORKSHOP

Authority check

if AUTH in {B, C} and route in {BOOKFREECALL, BOOKPAIDCALL}:
send CO_DECIDER template, then proceed when confirmed

Booking mutex

when generating [CALENDLYLINK]: lock slotid for 5 minutes; release on success/fail

Post‑booking

REMINDERS = ON (24h and 2h)
if PAID route and payment fails: retry at +1h and +6h; after 3rd fail → HUMAN
```

You need two cost lanes: messaging and model. Use conservative planning numbers and adjust to your BSP and region.

Messaging: All‑in per outbound message ≈ Meta template fee + BSP/platform markup. A practical US planning number is ~$0.03/message. If the chat starts from a Click‑to‑WhatsApp ad or CTA button, user‑initiated messaging is typically free within a limited window — run your qualifier + booking inside that window.

Model: Assume ~300 tokens/turn over ~6–8 turns → ~1,800–2,400 tokens/convo. On “mini” models, that’s low cents per convo; even full‑fat models are typically <$0.15 per convo when prompts are small.

Quick estimator (paste into a sheet):
```
Inputs: leadspermonth [100], engagerate [0.30], turnsperconvo [8], msgcost [0.03], tokensperturn [300], tokenpriceper1k [0.0025], adinitiated_share [0.50]

Derived:
convos = leadspermonth * engage_rate
messages = convos * turnsperconvo
paidmessages = messages * (1 - adinitiated_share)
wacost = paidmessages * msg_cost
llmtokens = convos turnsperconvo tokensper_turn
llmcost = (llmtokens/1000) * tokenpriceper_1k
totalopscost = wacost + llmcost
```
Example with defaults: convos=30; messages=240; paidmessages=120; wacost=$3.60; llmtokens=72,000; llmcost=$0.18; totalopscost ≈ $3.78. Twelve booked calls at 40% of engaged, three sales at $1.5k → $4,500 revenue on <$10 ops cost. Adjust to your geography/model.

Stay well inside policy so your number doesn’t get throttled while you’re in the air.

Consent: Always capture explicit opt‑in (surface, timestamp, exact language). Honor STOP/unsubscribe immediately.

Templates: Keep copy plain, specific, and non‑hype. Submit 2–3 variants per key step (greeting, booking, reminders). Avoid overusing “free,” vague CTAs, and promises.

Messaging limits: Ramp slowly. Week 1: ≤10 new convos/day; Week 2: ≤100/day if quality high; Week 3+: scale. Watch quality rating; reduce volume if it dips.

Data minimization: Store only what you need (consent, answers, booking/payment refs). Auto‑purge transcripts at 6 months; retain opt‑in proofs.

Task‑specific bot: Keep the assistant narrowly focused on qualification/booking — not a general chat AI.

Required footer cues: Periodically include “Reply STOP to opt out.” in non‑transactional messages.

Assume things will break. Design graceful failure.

Template rejections: Have alternates ready; strip promo phrasing; resubmit.

Hallucinations: Bot can only send pre‑approved templates; never compose free text. If message doesn’t match A/B/C/D or YES/STOP → trigger EDGECASE_HANDOFF and alert you.

Timezone misses: Let Calendly convert time zones; include local time echo in CONFIRM/REMINDER. Nightly manual “tomorrow’s calls” check.

Double bookings: Implement a 5‑minute booking lock (mutex) per slot while user is on the link.

Payment fails: Retry at +1h and +6h; offer alternative method (PayPal/Wise). After 3 fails → HUMAN.

Rate‑limit dips: Auto‑reduce daily start volume by 50% and swap to lower‑friction templates for 48h.

Monitoring: Real‑time Telegram alert on errors; daily digest: leads→convos→qualified→booked→show rate→costs. SLA: human reply within 12 hours on handoffs.

Use one of these positions; the tech stays the same.

1. $97 Audit Tripwire → $1.5k+ Implementation

Flow: PAID route with $97 deposit to book → 30‑min screen share audit → present 3 packages ($1.5k/$3k/$5k). Credit deposit toward purchase. Typical conversion: ~60% to at least the base package (your mileage will vary).

1. $49 Mini‑Workshop → $1k Cohort/Done‑For‑You

Flow: WORKSHOP route for Budget A or “just researching.” Live or pre‑record + WhatsApp Q&A. Follow‑up sequence pushes cohort/DFY.

1. Free Lead Magnet → $300 Strategy Call

Flow: NURTURE route delivers a 3‑day mini‑course in WhatsApp; Day 3 offers a $300 strategy session to filter tire‑kickers. Expect lower booking rate (~8%) but higher close quality.

Tip: Start with one archetype for 2 weeks, then A/B the deposit/workshop price — not the entire flow.