Automating complete business workflows

Section 10.6.4 : Automating complete business workflows

🎯 Learning objective

Design and implement end-to-end business automation workflows combining Make, AI, and your existing tools for measurable productivity gains. You'll move from automating isolated tasks to automating complete processes.

Think in "workflows", not "tasks"

The most common beginner mistake in automation: automating an isolated task. "I automated sending an email" — that's fine, but it's like replacing one link in the chain without touching the others.

The real ROI comes when you automate a complete process, end-to-end, where each step feeds the next without human intervention.

Comparison:

The method to move from task to workflow:

1. →Map the current process (who does what, when, with which tool)
2. →Identify repetitive steps and decision points
3. →Measure time spent on each step
4. →Design the automated workflow on paper (or with Mermaid)
5. →Build in phases: first the main path, then branches, then edge cases
6. →Test with real data (not just dummy data)
7. →Iterate for 2 weeks before moving to permanent production

Workflow 1: Automated client onboarding

Context: you're a freelancer/agency receiving 10-20 prospects per week. Each prospect must be qualified, recorded, sent a personalized email, and have a meeting scheduled. Manually: 30 min per prospect = 5-10h/week.

Make scenario details (6 modules + Router):

Module 1 — Typeform > Watch New Entries: The client fills out a qualification form with:

• →Name and email
• →Project type (website / mobile app / marketing / other)
• →Estimated budget (< $5K / $5-15K / $15-50K / > $50K)
• →Desired timeline (urgent < 1 month / normal 1-3 months / flexible > 3 months)
• →Needs description (free text)

Module 2 — OpenAI > Chat Completion: System prompt:

You are a business analyst. Analyze this prospect and return a JSON:
{
  "profile": "SMB" | "Startup" | "Corporate" | "Individual",
  "score": 1-10,  // lead quality based on budget, urgency, and clarity of need
  "identified_needs": ["need1", "need2", "need3"],
  "recommended_offer": "Starter Pack" | "Pro Pack" | "Enterprise Pack" | "Custom",
  "attention_points": "elements to discuss in the meeting",
  "personalized_email": "personalized welcome email of 150 words max",
  "priority": "high" | "normal" | "low"
}

Scoring rules:
- Budget > $15K AND urgency < 1 month = score 9-10
- Budget $5-15K = score 5-7
- Budget < $5K = score 2-4
- Vague description = -2 points

Module 3 — Router (3 routes):

• →Route A (score ≥ 8): priority prospect → VIP treatment
• →Route B (score 4-7): standard prospect → normal process
• →Route C (score < 4): cold prospect → polite email + add to newsletter

Module 4 — Google Sheets > Add Row (all routes): Records the lead with all fields + AI analysis in your lightweight CRM.

Module 5a (Route A) — Gmail > Send Email: Personalized VIP email + direct Calendly link with priority time slots.

Module 5b (Route B) — Gmail > Send Email: Standard email + Calendly link with regular time slots.

Module 5c (Route C) — Mailchimp > Add Subscriber: Added to newsletter with "cold prospect" tag.

Module 6 (Routes A and B) — Slack > Send Message: Notification to #new-leads: "🎯 New lead: {{name}} — AI Score: {{score}}/10 — {{recommended_offer}} — Priority: {{priority}}"

Measurable result: a process that took 30 min per prospect → now 100% automatic. For 15 prospects/week = 7.5 hours saved per week.

Workflow 2: Weekly content pipeline

This workflow automatically produces your weekly marketing content:

Detailed Make scenario:

📅 Monday 7am (Schedule trigger)
│
├─ Module 1: HTTP > Perplexity API
│  → "What are the 10 most important news items 
│     in [your industry] this week?"
│
├─ Module 2: OpenAI > Chat Completion
│  System: "Summarize the 10 news items. For each, provide:
│     title, 2-sentence summary, relevance 1-10, possible angle
│     for a LinkedIn post. JSON array."
│  User: {{1.response}}
│
├─ Module 3: Filter (relevance > 7)
│
├─ Module 4: OpenAI > Chat Completion
│  "For each filtered news item, write:
│   1. A professional LinkedIn post (200 words, with hook)
│   2. A Twitter thread of 4 tweets
│   3. A newsletter paragraph
│   Return as structured JSON."
│
├─ Module 5: Google Docs > Create Document
│  Title: "Content Week {{formatDate(now; 'W-YYYY')}}"
│  Content: formatted compilation of drafts
│
├─ Module 6: Slack > Send Message (#content-team)
│  "📝 Week's drafts ready: {{5.url}}
│   {{count}} LinkedIn posts, {{count}} Twitter threads
│   Review and publish before Wednesday."
│
└─ Module 7: Google Sheets > Add Row (tracking)
    Week | News Count | Content Count | Status

Advanced variant: add a second scenario that monitors the performance (likes, shares) of your published posts from the previous week and incorporates those insights into the following week's content generation.

Workflow 3: Continuous customer feedback analysis

📅 Daily 9am (Schedule) + Webhook (real-time for urgencies)
│
├─ Module 1: Typeform > Watch Responses (NPS/satisfaction survey)
│  + Google Sheets > Watch New Rows (manual feedback)
│
├─ Module 2: OpenAI > Chat Completion
│  System: "Analyze this customer feedback. Return a JSON:
│   {
│     sentiment: 'positive' | 'neutral' | 'negative' | 'very negative',
│     nps_score: estimate 0-10,
│     themes: ['theme1', 'theme2'],
│     key_quote: 'the most important sentence from the feedback',
│     urgency: true/false,
│     action: 'description of the recommended action',
│     department: 'support' | 'product' | 'sales' | 'management'
│   }"
│
├─ Module 3: Router
│  Route 1 (very negative OR urgency:true):
│    → Slack #client-alerts (mention @channel)
│    → Gmail → email to the relevant department manager
│    → Trello > Create Card (board "Urgencies")
│  Route 2 (negative, not urgent):
│    → Trello > Create Card (board "Improvements")
│    → Empathetic thank-you email to the client
│  Route 3 (positive):
│    → Google Sheets "Testimonials" (archive for marketing)
│    → Thank-you email + request Google/Trustpilot review
│
├─ Module 4 (all routes): Google Sheets > Add Row (full logging)
│
└─ 📅 Friday 5pm (second scenario — weekly report)
    → Google Sheets > Search Rows (week's feedback)
    → Aggregator
    → OpenAI > "Write a weekly satisfaction report:
       average NPS score, trends, top 3 issues, top 3 strengths,
       recommended actions for next week"
    → Gmail > Send (to leadership team)
    → Google Docs > Create Document (archive)

Workflow 4: Automated HR management (recruitment)

A workflow often overlooked but with very high ROI:

📧 Incoming email (careers@) → Trigger
│
├─ Module 1: Gmail > Watch Emails (label "Applications")
│
├─ Module 2: OpenAI > Chat Completion
│  "Analyze this application. Extract:
│   - name, email, target position
│   - estimated years of experience
│   - key skills identified
│   - fit with the position (score 1-10 + justification)
│   - interview questions to ask (3 suggestions)
│   - potential red flags
│   Attached resume: {{1.attachments[0].data}}"
│
├─ Module 3: Router
│  Route A (score ≥ 7): 
│    → Gmail: interview invitation email (Calendly link)
│    → Notion: create candidate profile
│    → Slack #recruiting: "✅ Promising profile: {{name}}"
│  Route B (score 4-6):
│    → Gmail: waiting email ("profile under evaluation")
│    → Notion: create "To review" profile
│  Route C (score < 4):
│    → Gmail: courteous and personalized rejection email
│    → Google Sheets: archive
│
└─ Final module: Google Sheets > Add Row (recruitment tracking)

Calculating the ROI of your automations

To convince your manager (or yourself) to invest time in automation, calculate the ROI precisely:

Automation ROI formula:

ROI = (Monthly gain — Monthly cost) / Setup cost × 100

Monthly gain = Time saved × Hourly rate + Errors avoided × Cost per error
Monthly cost = Make subscription + AI API costs + Maintenance (1h/month)
Setup cost = Design hours × Hourly rate

Concrete example — Onboarding workflow:

Monthly gain:
- 15 prospects/week × 30 min = 7.5h/week = 30h/month
- 30h × $50/h (loaded hourly rate) = $1,500/month
- Errors avoided (missed follow-ups): ~$200/month

Monthly cost:
- Make: $9 (Core plan, 10K ops)
- OpenAI API: ~$15 (gpt-4o-mini, 600 calls/month)
- Maintenance: 1h × $50 = $50
- Total: $74/month

Setup cost:
- Design + build + test: 8h × $50 = $400

Monthly ROI = ($1,700 - $74) / $400 × 100 = **406% per month**
Payback period: $400 / $1,626 = 0.25 months = **1 week**

ROI table for all 4 workflows: