Day 15 of 28 · AI Challenge
Automating Repetitive Work
⏱ 7 min
📊 Beginner
Welcome to Week 3 — this is where your AI skills go from "useful" to "unfair advantage." You've learned to write, research, and create with AI. Now you'll learn to make it work without you.
Today we're hunting down the tasks that steal your time and handing them to AI. The average knowledge worker spends 28% of their workday on email and 20% on data entry and searching for information. That's nearly half your day on autopilot work. Let's fix that.
The time audit
Before you automate anything, you need to know what's eating your hours. Here's a quick test: think about your last workday and ask yourself three questions.
What did I do more than twice? Formatting reports, copying data between tools, sending the same type of email, updating spreadsheets — anything repetitive is a candidate.
What did I dread doing? Dread usually means it's boring, repetitive, and low-value. Perfect automation material.
What fell through the cracks? Follow-ups you forgot, invoices you sent late, updates you never posted — if it slipped, it's because it should have been automatic.
Knowledge Check
Which of these tasks is the BEST candidate for AI automation?
A
Brainstorming a new product idea with your team
B
Categorizing and filing 50 incoming support emails every morning
C
Deciding whether to pivot your business model
D
Negotiating a major partnership deal
Categorizing emails is repetitive, rule-based, and high-volume — the perfect automation target. Strategic decisions like negotiations, brainstorming, and pivots still need human judgment and creativity.
Your automation toolkit
You don't need to write code to automate. Three categories of tools handle most business automation:
AI-native automation platforms — Tools like Zapier and Make (formerly Integromat) connect your apps and let AI make decisions in between. Zapier alone connects to 6,000+ apps. A basic plan starts around $20/month and can save you 10+ hours per week.
Built-in AI features — Gmail's Smart Compose, Notion AI, Excel Copilot, Slack AI summaries. You're probably already paying for tools that have automation features you've never turned on.
ChatGPT, Claude, and custom GPTs — For tasks that need language understanding: summarizing, categorizing, drafting, translating, extracting data from messy text.
Knowledge Check
What's the advantage of using tools like Zapier or Make over manually copying data between apps?
A
They make your apps run faster
B
They replace the need for any other software
C
They automatically move data and trigger actions between apps 24/7 without you lifting a finger
D
They're only useful for large enterprises
Automation platforms connect your existing tools and move data between them based on triggers you set. They work around the clock, never forget, and never make copy-paste errors. A freelancer saving 5 hours per week at $50/hour recovers $1,000/month — from a $20 tool.
Real example — auto-categorizing emails
Let's see this in action. Say you run a small consulting business and get 40-60 emails per day. Some are client inquiries, some are invoices, some are newsletters, and some are spam that got through. Manually sorting takes 30 minutes every morning.
I want to set up an automated workflow for my inbox. Every morning I get 40-60 emails and I waste 30 minutes sorting them. Here's what I need:
- Client emails → star them and move to "Clients" folder
- Invoices/receipts → move to "Finance" folder
- Newsletters → move to "Read Later" folder
- Everything else → leave in inbox
How do I set this up?
Great use case — here's exactly how to build this:
**Option 1: Zapier + AI (easiest)**
1. Create a Zap triggered by "New Email in Gmail"
2. Add a "ChatGPT" step that classifies the email into: Client, Finance, Newsletter, or Other
3. Add a "Filter" step that routes based on the classification
4. Add Gmail actions: apply label, move, or star
Total setup time: ~15 minutes
Cost: Zapier Starter ($20/mo) + your existing ChatGPT plan
**Option 2: Google Apps Script + AI (free but technical)**
I can write you a script that runs every hour, sends email subjects and senders to the OpenAI API, and sorts based on the response.
**Option 3: Gmail filters + AI labels (simplest)**
Use Gmail's built-in filters for obvious patterns (known client domains, invoice keywords) and only route ambiguous emails through AI.
I'd recommend starting with Option 3 for the easy wins, then adding Option 1 for the tricky emails. Want me to help set up either one?
Let's do Option 1. Walk me through the Zapier setup.
Here's your step-by-step:
**Step 1** — Go to Zapier → Create New Zap
**Trigger**: Gmail → "New Email" → Connect your account
**Step 2** — Add Action: "ChatGPT" → "Conversation"
Prompt: *"Classify this email. Subject: [Subject]. Sender: [From]. Reply ONLY with one word: Client, Finance, Newsletter, or Other."*
**Step 3** — Add Action: "Filter" → Only continue if ChatGPT output contains "Client"
**Step 4** — Add Action: Gmail → "Add Label" → Clients + "Star Email"
Duplicate Steps 3-4 for Finance and Newsletter labels.
That's it. Test it with 5-10 emails, tweak the prompt if needed, then let it run. You just got 30 minutes back every morning — that's **10+ hours per month**.
↻ Replay conversation
Knowledge Check
In the Zapier email workflow, what does the ChatGPT step actually do?
A
It rewrites the emails in a better tone
B
It replies to the emails automatically
C
It reads the email subject and sender, then classifies it into a category so Zapier knows where to route it
The AI step acts as the "brain" in the workflow — it reads the email metadata and makes a classification decision. Zapier then uses that decision to take the right action (label, move, star). The AI doesn't reply or modify the email — it just categorizes.
The 5-minute automation you can set up today
Don't wait for the "perfect" automation. Start with something small and immediate:
If you use Gmail: Set up 3-5 filters right now. Go to Settings → Filters → Create Filter. Route emails from your top clients to a priority folder. Route receipts (subject contains "receipt" or "invoice") to a finance folder. Takes 5 minutes, saves 5 minutes every day.
If you use Slack or Teams: Enable AI summaries on your busiest channels. Instead of reading 200 messages, you'll get a 30-second summary.
If you manage social media: Use a scheduling tool like Buffer or Later with AI-generated captions. Batch-create a week of posts in 30 minutes instead of posting daily.
The goal isn't to automate everything on Day 1. It's to automate one thing that you'll feel tomorrow morning.
Final Check
What's the smartest approach to getting started with automation?
A
Wait until you find the perfect tool before starting
B
Automate everything at once to maximize time savings
C
Only automate tasks that take more than 2 hours per day
D
Start with one small, repetitive task — prove it works, then build from there
Start small, prove value, then expand. One 5-minute automation that runs every day saves you 30+ hours per year. Stack a few of those and you've reclaimed entire workweeks. The key is momentum — your first automation gives you confidence to build the next one.
⚡
Day 15 Complete
"Every minute you spend on repetitive work is a minute you could spend on work that actually grows your business. Automate the boring stuff."
Tomorrow — Day 16
AI for Project Management
Tomorrow you'll learn to manage projects and teams with AI doing the heavy lifting.