How to Print $10K-50K/mo
With a $24/month AI Agent

npm install -g openclaw
openclaw onboard

Select Claude as your LLM. Connect your Telegram bot (create one via @BotFather on Telegram, takes 2 min).

Create a doc with:

• Every question customers ask (check your inbox, DMs, support tickets)
• The exact answer to each one
• Your policies (refunds, shipping, booking, pricing)
• Product/service details

Paste this into OpenClaw's knowledge base.

You are a customer support agent for [Business Name].
Your knowledge base contains everything you need to answer questions.

Rules:
- Be friendly but concise
- Answer questions using ONLY the knowledge base
- For order tracking: Pull status from [Shopify/CRM] API
- For appointment booking: Check availability in [Calendar] and confirm
- For refund requests: Verify purchase date, if under 30 days approve automatically
- For anything outside your knowledge base: Say "Let me connect you
  with the team" and alert via Telegram
- Never make up information
- Never promise something not in your knowledge base

alert_rules:
  - trigger: "cannot answer"
    action: telegram_notify
    message: "Customer needs human help: {customer_message}"
  - trigger: "refund over $100"
    action: telegram_notify
    message: "Large refund request: {details}"

You only get pinged for edge cases. Everything else runs automatically.

Send 20 test messages covering common questions (should answer perfectly), edge cases (should escalate), and weird questions (should not hallucinate). Fix any responses that sound off. Tighten the prompt.

• Let it run for 48 hours — check every response
• After 48 hours — check twice daily
• After 1 week — check once daily
• After 1 month — check every few days

Before building anything, answer: What industry? What job title? What company size? What location? What triggers buying (hiring, funding, expansion)?

You are a lead qualification agent for [Business].
For each lead, research:
- Company website (what they do, how big)
- Recent news (funding, hiring, expansion)
- LinkedIn activity (are they active, what do they post about)
- Tech stack if relevant (BuiltWith data)

Score 1-10 based on:
- Budget likelihood (company size, funding)
- Need likelihood (hiring, growth signals)
- Timing (recent triggers)

Output: Lead name, score, reasoning, recommended approach.
Only pass leads scoring 7+ to outreach.

Connect via n8n: Lead scores 7+ → Add to outreach sequence. Personalized email via SendGrid. Follow-up scheduled for Day 3 and Day 7.

Write a cold email to [Lead Name] at [Company].
Context: [What they do, recent trigger, why they might need us].
Offer: [Your service].
Tone: Casual, not salesy, like a human wrote it.
Length: Under 100 words.
CTA: Soft ask for a quick call.

Do NOT use: "I hope this email finds you well",
"Just following up", "Pick your brain",
or any generic opener.

• Positive reply → Alert you on Telegram immediately, suggest meeting times
• Question → Answer from knowledge base, keep conversation going
• Not interested → Mark in CRM, do not follow up
• Out of office → Reschedule follow-up

End-of-day cron job: Leads scraped, leads qualified (7+ score), emails sent, responses received, meetings booked. Summary to Telegram.

In n8n, create a workflow: Trigger daily at 7am → Apify Facebook Ads Library scraper (niche keywords + competitor brands) → JSON output with creative, copy, engagement metrics. Same for TikTok.

Analyze these ads from [niche]. For each ad, extract:
- Hook (first 3 seconds or first line)
- Angle (what pain/desire it targets)
- Format (UGC, static, animation, etc.)
- Engagement signals (likes, comments, shares)

Then:
- Rank top 10 by engagement
- Identify common patterns in winners
- Flag hooks that appear multiple times (proven)
- Note what's missing that we could test

Output as a table.

Take the winning hooks and multiply them: 50 variations of the #1 hook, 30 of the #2 hook, 20 completely new angles. Each with hook, body copy (2-3 sentences), and CTA. Human-sounding, no corporate speak.

Connect ElevenLabs API. Convert top scripts to voiceover — casual, energetic, natural pacing. Now you have audio for video ads without hiring anyone.

Connect to Facebook Ads Manager API via n8n. Take top 20 variations, create ad set with $5-10/variation budget, let run 48-72 hours, pull performance, analyze winners and losers with AI, generate next round of iterations.

What research do you do repeatedly? Competitor analysis (pricing, features, positioning), market sizing (TAM, trends), company research (background, funding, team), content research (what's ranking, what angles work). Create a template for each.

You are a research analyst. Task: [Competitor Analysis].
For company: [Company Name]. Research and compile:
- What they sell (products/services, pricing)
- Who they target (customer segments, use cases)
- How they position (messaging, unique angles)
- Their content (blog topics, social presence, ad messaging)
- Weaknesses (negative reviews, gaps in offering)
- Recent news (funding, launches, changes)

Sources: Company website, LinkedIn, G2/Capterra reviews,
Crunchbase, Perplexity search.

Output: Structured report with sources cited. 1-2 pages.

"Research these 10 competitors. For each, compile the standard competitor analysis. Output as Google Sheet with one row per company." The agent does 10 companies in the time you'd do 1.

Schedule weekly market reports: top news, competitor moves, trends, recommended actions. Client gets consistent value. You spend 10 minutes reviewing instead of 3 hours writing.

Search: "Shopify store owner" OR "real estate agent" OR "agency founder"

Hey [Name], saw you run [Business]. Quick question —
how much time does your team spend on [support tickets /
lead follow-up / manual research] per week?

I've been helping businesses automate that with AI agents.
One client cut their support time by 70%.

Worth a 10 min call to see if it'd work for you?

Volume: 50 DMs/day. Expect 5-10% reply rate. 1-2 calls/day.

Create gig: "I'll build a custom AI agent that automates [task]." Price low to start ($200-500) for reviews. 10+ reviews → raise to $1,000+.

Walk into dentists, salons, gyms, restaurants. One local client at $500/mo = $6K/year. Get 10 and that's $60K from local alone.