5 Boring AI Automations That Can Quietly Print $2K

Life of Arjav Playbook
Topic: practical n8n automation setups you can build, package, and sell to service businesses, creators, agencies, consultants, and local operators.

These are not flashy “AI agent” demos. They are small operating systems for repetitive business work: lead capture, content repurposing, invoice logging, review replies, and support ticket routing.

---

Quick Positioning

Most businesses do not need a custom SaaS product. They need a workflow that removes a recurring operational problem.

That is the offer:

“I will set up a working automation that removes this manual task from your week, connects to your existing tools, and gives you a clean approval step before anything customer-facing goes out.”

The 5 Setups

---

Stack Assumptions

These setups are built around n8n because it supports triggers, HTTP requests, Google Workspace, Gmail, Google Sheets, webhooks, branching, AI nodes, and third-party APIs.

Use equivalent tools if your client already runs Make, Zapier, Pipedream, Relay, Bardeen, Airtable Automations, or custom scripts.

Core tools

• n8n: workflow orchestration
• Google Sheets or Airtable: lightweight database
• Gmail / Google Workspace: inbox triggers and send actions
• HTTP Request node: API calls to external systems
• AI model node or HTTP call: classification, summarization, drafting
• Slack or email: approval and alerts

Optional tools

• Apollo API: people search and lead data enrichment
• Apify: scraping actors for public web sources such as Google Maps or Meta Ad Library
• Smartlead or Instantly: outbound campaign delivery
• Deepgram / AssemblyAI / OpenAI transcription: audio/video transcription
• Buffer: social scheduling
• QuickBooks Online: invoice creation or accounting sync
• Zendesk / Freshdesk / HelpScout: support ticketing
• Google Business Profile API / Trustpilot API: review workflows

---

Before You Build: The Client Intake Checklist

Use this before touching the workflow.

Access

• n8n workspace or permission to create one
• Google account access for Sheets, Drive, Gmail
• API keys for Apollo, Apify, Smartlead, Instantly, Buffer, QuickBooks, Zendesk, or other client tools
• Client brand voice examples: website, emails, replies, social captions, support macros
• Approval channel: Slack, Gmail, Notion, Airtable, or Google Sheet

Rules

• What should run automatically?
• What needs human approval?
• What should never be sent automatically?
• What fields are required before a record moves forward?
• What happens when confidence is low?
• Who owns failed workflow runs?

Minimum guardrails

• Never send cold outreach without deduplication and suppression checks.
• Never auto-post review replies for 1-star or legal-sensitive reviews unless the client explicitly approves.
• Never auto-send refund, cancellation, pricing, or contract replies without approval.
• Log every external action: sent email, created ticket, posted reply, added lead, created invoice.

---

Setup 1: Lead Scraper

Promise

It finds leads while you sleep.

It scrapes or pulls fresh leads, enriches them, filters them against your ICP, dedupes them, and sends approved leads into an outreach campaign or CRM.

Best-fit clients

• B2B agencies
• recruiters
• consultants
• local lead generation operators
• outbound teams
• niche data providers
• appointment setters

What to sell

“I’ll build you a daily lead engine that finds new prospects, enriches them, checks if they fit, dedupes against your sheet or CRM, and pushes only usable leads into your outreach system.”

Data sources

Pick one source first. Do not overbuild.

Option A: Apollo API

Use when the client has a defined B2B ICP.

Useful filters:

• person titles
• seniority
• company headcount
• company geography
• industry
• keywords
• company domain

Apollo’s People API supports search by person and organization filters, including job title and location. Source: Apollo People API Search.

Option B: Apify Google Maps Scraper

Use for local business lead generation.

Good niches:

• clinics
• gyms
• med spas
• dentists
• roofers
• restaurants
• hotels
• wedding venues
• real estate agencies

Apify actors can be started via API, and their run output can be fetched from the default dataset after the run completes. Source: Apify Run Actors.

Option C: Manual CSV seed list

Use when the client already has a list but it is messy.

The workflow cleans, dedupes, enriches, scores, and pushes only usable rows forward.

---

Recommended n8n Flow

Schedule Trigger
  → HTTP Request: Apollo / Apify / Scraper
  → Code: normalize fields
  → Google Sheets: lookup existing leads
  → IF: skip duplicates
  → AI Classifier: ICP fit score
  → IF: only score >= threshold
  → Email verification / enrichment API
  → Google Sheets: append approved lead
  → Smartlead / Instantly: add lead to campaign
  → Slack / Gmail: daily summary

Node-by-node build

1. Schedule Trigger

Run once per day or every weekday morning.

Recommended setting:

Every weekday at 07:00 client timezone

n8n’s Schedule Trigger is designed for recurring workflows. Source: n8n Schedule Trigger node.

2. HTTP Request: Lead source

Use the HTTP Request node for Apollo, Apify, or a custom data source. n8n’s HTTP Request node can be configured from API docs or curl examples. Source: n8n HTTP Request node.

Example Apollo-style payload:

{
  "person_titles": ["Founder", "CEO", "Head of Marketing"],
  "person_locations": ["United States"],
  "organization_num_employees_ranges": ["11,50", "51,200"],
  "page": 1,
  "per_page": 25
}

3. Normalize fields

Use a Code node to standardize the output.

Target schema:

{
  "first_name": "",
  "last_name": "",
  "email": "",
  "title": "",
  "company_name": "",
  "company_domain": "",
  "linkedin_url": "",
  "source": "",
  "city": "",
  "country": "",
  "icp_score": 0,
  "status": "new"
}

4. Deduplication

Deduplicate against:

• email
• company domain
• LinkedIn URL
• company name + title combination

Store dedupe keys in a sheet column:

{{ $json.email.toLowerCase().trim() }}
{{ $json.company_domain.toLowerCase().replace('www.', '') }}

5. ICP classifier

Prompt template:

You are scoring leads for a B2B service provider.

Return strict JSON only:
{
  "score": 0-100,
  "fit": "high" | "medium" | "low" | "exclude",
  "reason": "short reason",
  "missing_fields": []
}

ICP:
- Target industries: [INSERT]
- Target company size: [INSERT]
- Target geos: [INSERT]
- Bad fits: [INSERT]

Lead:
{{ JSON.stringify($json) }}

6. Push to outreach

Smartlead supports adding leads to campaigns by API. Source: Smartlead Add Leads to Campaign.

Instantly supports creating leads and bulk adding leads to a campaign/list by API. Source: Instantly Lead API.

Only push leads after:

• valid email exists
• duplicate check passed
• suppression check passed
• ICP score passes threshold
• source is logged

Sheet structure

Human approval version

Add a Google Sheet column:

approved_for_outreach = yes/no

Then add an IF node before Smartlead/Instantly:

Only continue if approved_for_outreach = yes

Pricing

Common mistakes

• Pulling too many leads before proving one niche works.
• Sending leads to outreach before suppression and dedupe.
• Not logging skipped records.
• Using scraped data without checking the client’s legal and platform obligations.
• Selling “leads” when the real value is qualified pipeline.

---

Setup 2: Repurposer

Promise

One video becomes twenty posts.

Drop in a long video, podcast, webinar, Zoom recording, or Loom. The workflow transcribes it, extracts ideas, writes captions, creates short-form clips or post drafts, and queues everything for approval.

Best-fit clients

• creators
• coaches
• consultants
• personal brands
• agencies
• founder-led SaaS companies
• podcast hosts
• YouTubers

What to sell

“I’ll build a content repurposing engine that turns one long-form recording into clips, captions, carousel outlines, LinkedIn posts, X threads, and a weekly approval board.”

Recommended n8n Flow

Google Drive Trigger
  → Download file
  → Transcription API
  → AI: extract hooks and ideas
  → AI: write platform-native drafts
  → Google Sheets / Airtable: content queue
  → Slack / Gmail: approval request
  → Buffer: schedule approved posts

The Google Drive Trigger can start workflows when files are created or changed in Drive. Source: n8n Google Drive Trigger.

Buffer’s API supports creating and scheduling posts across major social platforms. Source: Buffer Posts and Scheduling.

Folder setup

Create this Drive structure:

/Content Engine
  /01 Incoming Videos
  /02 Transcripts
  /03 Drafts
  /04 Approved
  /05 Scheduled
  /06 Published

Content output map

AI extraction prompt

Analyze this transcript and extract reusable content assets.

Return strict JSON only:
{
  "main_idea": "",
  "audience": "",
  "best_hooks": [],
  "clip_candidates": [
    {
      "start_time": "",
      "end_time": "",
      "reason": "",
      "hook": ""
    }
  ],
  "linkedin_posts": [],
  "instagram_carousel_ideas": [],
  "x_thread_ideas": [],
  "newsletter_angle": ""
}

Transcript:
{{ $json.transcript }}

Caption writing prompt

Write platform-native captions from this idea.

Rules:
- no hype
- no generic motivational filler
- short hook first
- clear body
- strong CTA
- write like a practical operator
- avoid AI tells

Brand voice:
[PASTE 5-10 examples]

Idea:
{{ $json.idea }}

Approval board fields

Optional clip automation

For video clipping, keep it semi-automated first.

Recommended first version:

1. AI identifies clip timestamps.
2. Editor or VA checks the timestamps.
3. Clips are cut manually or with a clipping tool.
4. Captions and post copy are automated.

Fully automated clipping is possible, but it breaks more often than copy generation because framing, subtitles, silence removal, and speaker cuts need quality control.

Pricing

Common mistakes

• Publishing without approval.
• Writing every platform in the same tone.
• Not storing reusable hooks and angles.
• Trying to fully automate video editing before proving the copy workflow.
• Not building a content calendar view.

---

Setup 3: Invoice Autopilot

Promise

It does your invoices for you.

The workflow watches email and/or Drive for invoices, extracts invoice fields, logs them into a sheet, flags overdue items, and optionally creates or updates records in QuickBooks.

Best-fit clients

• agencies
• ecommerce businesses
• accountants
• bookkeepers
• creators with contractors
• operators handling recurring vendor invoices
• SMBs with messy inbox accounting

What to sell

“I’ll build an invoice autopilot that reads incoming invoice emails and PDFs, extracts the numbers, logs everything into a finance tracker, and flags missing or overdue payments before they become a problem.”

Recommended n8n Flow

Gmail Trigger: invoice email received
  → IF: has attachment or invoice keywords
  → Extract attachment / email body
  → OCR or document extraction API
  → AI: normalize invoice fields
  → Google Sheets: append row
  → IF: overdue or duplicate
  → Gmail / Slack: send alert
  → Optional: QuickBooks Online create/update invoice

The Gmail Trigger can return email metadata and full email content. Source: n8n Gmail Trigger.

Google Sheets in n8n supports document and sheet operations for storing structured workflow data. Source: n8n Google Sheets node.

QuickBooks Online exposes accounting API operations including invoice creation through its Accounting API. Source: QuickBooks Online Accounting API Postman collection.

Gmail search filters

Start with a label or search query:

subject:(invoice OR receipt OR payment due OR statement) has:attachment newer_than:7d

Create a Gmail label:

Finance / Incoming Invoices

Only trigger on that label if the inbox is noisy.

Invoice extraction schema

{
  "vendor_name": "",
  "vendor_email": "",
  "invoice_number": "",
  "invoice_date": "",
  "due_date": "",
  "currency": "",
  "subtotal": 0,
  "tax": 0,
  "total": 0,
  "line_items": [],
  "payment_terms": "",
  "bank_details_present": true,
  "confidence": 0,
  "source_email_id": "",
  "attachment_url": ""
}

Extraction prompt

Extract invoice data from the text below.

Return strict JSON only. Do not guess. If a field is missing, return null.

Required fields:
- vendor_name
- invoice_number
- invoice_date
- due_date
- currency
- subtotal
- tax
- total
- payment_terms
- confidence from 0 to 100

Invoice text:
{{ $json.invoice_text }}

Sheet structure

Logic rules

Duplicate rule

Mark as duplicate if:

vendor_name + invoice_number + total

already exists.

Needs review rule

Send to review if:

• confidence is below 85
• invoice number is missing
• total is missing
• currency is missing
• due date is missing
• vendor is new
• total is above client approval threshold

Overdue rule

Run a second daily workflow:

Schedule Trigger
  → Google Sheets: read unpaid invoices
  → IF: today > due_date
  → Slack/Gmail alert

Pricing

Common mistakes

• Treating OCR output as reliable without confidence scoring.
• Not preserving the original invoice attachment.
• Not building a duplicate rule.
• Auto-paying invoices. Do not do this unless the client has a formal approval process.
• Mixing receivables and payables in the same sheet without a type field.

---

Setup 4: Review Responder

Promise

It replies to reviews in your voice.

A new review lands. The workflow classifies the sentiment, drafts a brand-safe response, routes risky reviews for approval, and optionally posts approved replies.

Best-fit clients

• local businesses
• hotels
• restaurants
• clinics
• ecommerce brands
• agencies managing reputation for clients
• franchises
• service businesses

What to sell

“I’ll build a review response system that watches your reviews, drafts replies in your brand voice, escalates negative reviews, and gives your team an approval button before anything public goes live.”

Review sources

Google Business Profile

The Google Business Profile API supports working with review data, including listing reviews and replying to reviews. Source: Google Business Profile review data.

Trustpilot

Trustpilot offers APIs and webhooks for review-related workflows. Trustpilot’s help center describes webhooks for events such as new, deleted, or revised reviews. Source: Trustpilot webhooks.

Social reviews and mentions

Use platform-specific APIs where available. For unavailable sources, use a monitoring tool, email notifications, or a webhook-based scraper with permission.

Recommended n8n Flow

Webhook or Schedule Trigger
  → HTTP Request: fetch new reviews
  → Google Sheets: check review_id dedupe
  → AI: classify sentiment and risk
  → AI: draft reply in brand voice
  → IF: negative/legal/high-risk → approval required
  → Slack/Gmail: send approval request
  → IF approved → HTTP Request: post reply
  → Google Sheets: log final status

Sentiment classifier prompt

Classify this customer review.

Return strict JSON only:
{
  "sentiment": "positive" | "neutral" | "negative",
  "risk_level": "low" | "medium" | "high",
  "requires_human_approval": true,
  "reason": "",
  "recommended_tone": ""
}

High risk includes legal threats, refunds, safety issues, medical claims, discrimination, harassment, fraud, public accusations, or angry 1-star reviews.

Review:
{{ $json.review_text }}

Reply writer prompt

Write a public review reply for this business.

Rules:
- sound human and specific
- do not over-apologize
- do not mention internal policy
- do not offer refunds publicly
- do not argue with the reviewer
- keep it under 70 words
- invite private follow-up only when useful
- match the brand voice examples below

Brand voice examples:
[PASTE 5-10 APPROVED REVIEW REPLIES]

Review:
{{ $json.review_text }}

Classification:
{{ JSON.stringify($json.classification) }}

Auto-post rules

Safe to auto-post only when:

• sentiment is positive or neutral
• no legal/safety/medical/refund issue exists
• review is 4 or 5 stars
• confidence is high
• client has approved auto-posting

Always require approval when:

• review is 1 or 2 stars
• refund is mentioned
• legal threat is mentioned
• employee misconduct is mentioned
• medical or safety claim is mentioned
• customer names a specific staff member negatively

Sheet structure

Pricing

Common mistakes

• Auto-posting negative review replies without approval.
• Replying with generic templates.
• Not logging final replies.
• Ignoring revised or deleted reviews.
• Not adapting tone by review type.

---

Setup 5: Ticket Triage

Promise

It sorts every support ticket.

Every incoming support email or ticket gets classified, tagged, routed to the right owner, and answered automatically when the issue is simple and approved.

Best-fit clients

• ecommerce brands
• SaaS companies
• agencies
• course sellers
• marketplaces
• service businesses
• customer support teams

What to sell

“I’ll build a support triage system that reads incoming tickets, classifies the issue, assigns the right priority and owner, drafts replies, and automatically answers the low-risk repetitive tickets.”

Ticket sources

• Gmail shared inbox
• Zendesk
• HelpScout
• Freshdesk
• Intercom
• Front
• Tidio
• Crisp
• Typeform support form

Zendesk’s Ticketing API supports ticket creation, updates, comments, status changes, and bulk operations. Source: Zendesk Tickets API.

n8n’s Switch node is useful for branching workflows into different paths based on conditions. Source: n8n Switch node.

Recommended n8n Flow

Gmail Trigger / Zendesk Trigger / Webhook
  → AI: classify ticket
  → Switch: route by category
  → IF: low-risk FAQ → draft or send reply
  → IF: medium-risk → assign owner + draft reply
  → IF: high-risk → urgent escalation
  → Zendesk/Gmail: update ticket
  → Google Sheets: log handling metrics

Classification taxonomy

Start with 8 categories.

Classifier prompt

Classify this support ticket.

Return strict JSON only:
{
  "category": "order_status" | "refund_request" | "cancellation" | "billing_issue" | "technical_issue" | "product_question" | "complaint" | "partnership_sales" | "other",
  "priority": "low" | "medium" | "high" | "urgent",
  "customer_intent": "",
  "requires_human": true,
  "suggested_owner": "support" | "billing" | "technical" | "sales" | "founder",
  "confidence": 0,
  "reason": ""
}

Ticket:
{{ $json.ticket_text }}

Reply drafting prompt

Draft a support reply.

Rules:
- answer directly
- be concise
- do not promise refunds, discounts, or policy exceptions
- ask for missing information only if needed
- match the company tone
- keep under 120 words

Company policy:
[PASTE POLICY]

Customer ticket:
{{ $json.ticket_text }}

Classification:
{{ JSON.stringify($json.classification) }}

Routing rules

Auto-answer candidates

• order tracking link found
• product FAQ exists
• password reset instructions
• delivery ETA request
• return policy explanation without special case
• simple business hours or location question

Human approval required

• refunds
• cancellations
• chargebacks
• legal threats
• angry complaints
• enterprise customers
• subscription billing issues
• anything below confidence threshold

Sheet structure

Pricing

Common mistakes

• Auto-answering refund or billing issues.
• Not loading the company policy into the workflow.
• Using only one broad “support” category.
• Not measuring first response time and resolution time.
• Not routing VIP or angry customers differently.

---

Client Delivery Template

Use this structure when delivering any of the 5 systems.

1. Workflow map

Include a simple diagram:

Trigger → Collect → Clean → Classify → Approve → Act → Log → Report

2. Credentials checklist

3. Test cases

Each setup needs at least 10 realistic test cases.

Example for Ticket Triage:

4. Error handling

Every workflow should include:

• error branch
• failed API call logging
• retry strategy
• daily summary
• owner notification
• manual override field

5. Client handover

Deliver these files:

• n8n exported workflow JSON
• Loom walkthrough
• credentials map
• Google Sheet database
• SOP for approvals
• test case sheet
• maintenance notes

---

Maintenance Plan

Sell maintenance separately.

Maintenance is where these setups become profitable. The initial build pays for implementation. The monthly plan pays for keeping the system alive.

---

Sales Angles

Angle 1: “Admin leak”

“You are not losing time to strategy. You are losing time to tiny repeated admin loops. This removes one of them permanently.”

Angle 2: “Approval-first automation”

“This does not blindly send things on your behalf. It drafts, classifies, logs, and waits for approval where the risk is high.”

Angle 3: “Not a SaaS migration”

“No new platform for the team to learn. This connects to the tools you already use.”

Angle 4: “Built around your actual work”

“Most automation fails because it is built around a demo, not your messy process. This starts with your inbox, files, tickets, and real examples.”

---

Outreach Script

Cold email version

Subject: quick workflow idea for {{company_name}}

Hey {{first_name}},

I noticed {{company_name}} has a lot of repetitive operational work around {{specific_area}}.

I build practical AI automations that plug into tools teams already use: Gmail, Sheets, Slack, n8n, Zendesk, QuickBooks, Smartlead, Instantly, and similar.

One relevant idea for you:

{{workflow_name}}
It would {{one_line_outcome}}.

The useful part is that it can keep a human approval step for anything sensitive, so it does not become one of those reckless “AI agent” setups.

Worth a quick look?

Best,
{{your_name}}

DM version

Saw a workflow gap you could probably automate:

{{workflow_name}}

It would {{one_line_outcome}} using the tools you already have. No new SaaS migration. Just n8n + your current stack + approval steps where needed.

Want me to send the rough workflow map?

---

Build Priority

Do not build all five at once.

Start here:

1. Ticket Triage if the client has daily support volume.
2. Invoice Autopilot if the client has messy finance admin.
3. Lead Scraper if the client already sells through outbound.
4. Repurposer if the client publishes consistently.
5. Review Responder if the client has public reviews and local reputation risk.

Best first build:

Ticket Triage or Invoice Autopilot.

They are less dependent on creative quality and easier to prove with before/after metrics.

---

Source Notes

These references were used to keep the setups technically grounded:

• n8n HTTP Request node: https://github.com/n8n-io/n8n-docs/blob/main/docs/integrations/builtin/core-nodes/n8n-nodes-base.httprequest/README.md
• n8n Schedule Trigger: https://github.com/n8n-io/n8n-docs/blob/main/docs/integrations/builtin/core-nodes/n8n-nodes-base.scheduletrigger/README.md
• n8n Gmail Trigger: https://github.com/n8n-io/n8n-docs/blob/main/docs/integrations/builtin/trigger-nodes/n8n-nodes-base.gmailtrigger/README.md
• n8n Google Drive Trigger: https://github.com/n8n-io/n8n-docs/blob/main/docs/integrations/builtin/trigger-nodes/n8n-nodes-base.googledrivetrigger/README.md
• n8n Google Sheets node: https://github.com/n8n-io/n8n-docs/blob/main/docs/integrations/builtin/app-nodes/n8n-nodes-base.googlesheets/README.md
• n8n Switch node: https://github.com/n8n-io/n8n-docs/blob/main/docs/integrations/builtin/core-nodes/n8n-nodes-base.switch.md
• Apollo People API: https://docs.apollo.io/reference/people-api-search
• Apify Run Actors: https://docs.apify.com/actors/running
• Smartlead Add Leads to Campaign: https://api.smartlead.ai/api-reference/campaigns/add-leads
• Instantly Lead API: https://developer.instantly.ai/api-reference/groups/lead
• Buffer Posts and Scheduling: https://developers.buffer.com/guides/posts-and-scheduling.html
• Google Business Profile review data: https://developers.google.com/my-business/content/review-data
• Trustpilot webhooks: https://trustpilot.zendesk.com/hc/en-us/sections/4403187805586-Webhooks
• Zendesk Tickets API: https://developer.zendesk.com/api-reference/ticketing/tickets/tickets/
• QuickBooks Online Accounting API: https://www.postman.com/intuit-developer/intuit-developer-quickbooks-online-accounting-api/documentation/4884662-e6c576f1-f6d3-440f-b090-da9ff1ac519d

---

Final CTA

Comment ACCESS to get the guides, repos, workflows, and build breakdowns.

Save this if you are building client automations.

Follow @lifeofarjav for practical AI systems, automation workflows, and operator-grade implementation guides.