Automating Repetitive Work
Find your 'time vampires' and eliminate them — turning 2 hours of busywork into 5 minutes.
Marcus was building the same report every Monday morning for two years
Marcus is a marketing manager at a SaaS company. Every Monday at 8 AM, he opens five browser tabs: Google Analytics, HubSpot, the ad dashboard, the social media dashboard, and a shared Google Sheet. He copies numbers from each one. He pastes them into a slide deck template. He calculates week-over-week changes by hand. He writes a three-paragraph summary.
It takes three hours. Every. Single. Week.
One Tuesday, Marcus stayed late and spent 90 minutes setting up an AI-powered workflow. He built a prompt template that takes raw numbers and generates the summary. He set up a simple automation that pulls the data into one place. He created a prompt that formats the slide deck text.
The next Monday, the report took ten minutes. Marcus spent the other two hours and fifty minutes on strategy work his boss had been asking about for months.
That three hours wasn't just time — it was Marcus's best morning energy, burned on copy-paste. AI gave it back.
Time vampires: the tasks that drain your week without adding value
A time vampire is any task that:
- You do repeatedly (weekly, daily, or multiple times per day)
- Follows roughly the same steps every time
- Doesn't require deep creative thinking or complex judgment
- Would take the same amount of time whether it's your first year or your tenth
Think of it like a dishwasher. Before dishwashers, you spent 30 minutes after every meal scrubbing plates. The dishwasher didn't eliminate the work entirely — you still load it, add soap, unload it. But the actual scrubbing? The repetitive, low-skill, time-consuming part? That runs without you.
AI automation works the same way. You still set it up, review the output, and make decisions. But the repetitive middle part — the scrubbing — happens without you.
The time vampire identification matrix
| Question | If yes... |
|---|---|
| Do I do this task at least once a week? | Potential vampire |
| Does it follow the same steps every time? | Strong vampire signal |
| Could I write instructions for someone else to do it? | Definitely automatable |
| Does it involve copying data from one place to another? | Classic vampire |
| Am I essentially filling in a template? | Vampire confirmed |
| Does it require me to read and summarize something? | AI handles this well |
There Are No Dumb Questions
"But my repetitive tasks have nuances — every report is slightly different."
That's the beauty of AI automation vs. old-school automation. Traditional automation (macros, scripts) breaks when the input changes. AI handles variation. Your weekly report has different numbers each week, different highlights, different context — AI doesn't care. It adapts to the input. The structure is the same; the content changes. That's exactly what AI is built for.
"Isn't this just macros with extra steps?"
Macros follow rigid instructions: "copy cell A1 to B1." If the data moves to A2, the macro breaks. AI understands intent: "find the revenue number and put it in the summary." If revenue moves from row 5 to row 7, or the column header changes from "Rev" to "Revenue," AI still finds it. That's the difference between automation and intelligent automation.
Find your time vampires
25 XPThree levels of AI automation
You don't have to build a complex system. Start simple and level up.
Level 1: Template-based automation (start here)
You create a reusable prompt template with blanks that change each time.
Example — weekly report summary:
You are a marketing analyst writing a weekly performance summary.
Here are this week's numbers:
- Website visits: [PASTE NUMBER]
- New leads: [PASTE NUMBER]
- Conversion rate: [PASTE NUMBER]
- Ad spend: [PASTE NUMBER]
- Revenue attributed: [PASTE NUMBER]
Last week's numbers for comparison:
- Website visits: [PASTE NUMBER]
- New leads: [PASTE NUMBER]
- Conversion rate: [PASTE NUMBER]
- Ad spend: [PASTE NUMBER]
- Revenue attributed: [PASTE NUMBER]
Write a 3-paragraph executive summary highlighting:
1. The biggest win this week
2. The biggest concern
3. One recommended action for next week
Use a professional but conversational tone. Include the
week-over-week percentage change for each metric.
Save this template. Every Monday, paste in the new numbers. Done in 2 minutes instead of 45.
Level 2: Batch processing (handle 100 tasks at once)
Instead of running one thing through AI at a time, you batch them.
Examples of batch processing:
| Single task (slow) | Batch version (fast) |
|---|---|
| Summarize one meeting transcript | "Here are 5 meeting transcripts. Summarize each one in 3 bullet points." |
| Write one product description | "Here are 20 products with their specs. Write a 50-word description for each." |
| Classify one support ticket | "Here are 50 support tickets. Categorize each as: billing, technical, feature request, or other." |
| Personalize one email | "Here are 15 customer names and their industries. Write a personalized opening line for each." |
The key insight: AI doesn't get tired. Processing one item or fifty items takes roughly the same mental effort from you — you just paste more data.
Level 3: Simple workflows ("When X happens, AI does Y")
This is where you connect AI to triggers so it runs automatically.
Tools that make this easy (no code required):
| Tool | What it does | Example workflow |
|---|---|---|
| Zapier + AI | Connects apps with AI steps | New form submission → AI categorizes → routes to correct Slack channel |
| Microsoft Power Automate + Copilot | Automates Office 365 workflows | Email received → AI summarizes → adds to weekly digest |
| Google Apps Script + Gemini | Automates Google Workspace | New spreadsheet row → AI generates report → emails to manager |
| Make (formerly Integromat) | Visual workflow builder | Support ticket created → AI drafts response → queues for review |
AI Automation Pipeline
There Are No Dumb Questions
"Do I need to learn to code for Level 3?"
No. Tools like Zapier and Make are visual — you drag and drop steps. The AI piece is usually "add an AI step" where you paste a prompt template. If you can write a prompt, you can build these workflows. That said, having a tech-savvy colleague review your first workflow is smart.
"What if the AI makes a mistake in an automated workflow?"
Build in a "human review" step for anything important. The automation runs, AI does its thing, but before the final output is sent/posted/submitted, it lands in your inbox for a 30-second review. This catches errors while still saving you 90% of the time.
Design your first automation
50 XPEstimating your time savings
Here's a simple formula to calculate the ROI of automating a task:
Time saved per month =
(Manual time per task × Frequency per month)
− (Automated time per task × Frequency per month)
− Setup time amortized over 6 months
Realistic time savings by automation level
| Automation level | Typical setup time | Typical time savings | Best for |
|---|---|---|---|
| Level 1: Templates | 15-30 min once | 50-70% per task | Reports, summaries, standard communications |
| Level 2: Batch processing | 30-60 min to learn | 80-90% per batch | Bulk content, data classification, personalization |
| Level 3: Workflows | 2-4 hours once | 90-95% ongoing | Recurring processes, routing, monitoring |
The automation priority matrix
Not every time vampire deserves automation. Use this to prioritize:
- Top-right (high frequency, high time): Automate first. This is where Marcus's Monday report lived.
- Top-left (low frequency, high time): Automate if the setup is easy. Quarterly reports fit here.
- Bottom-right (high frequency, low time): Consider templates. Quick replies, status updates.
- Bottom-left (low frequency, low time): Don't bother automating. The setup cost exceeds the savings.
✗ Without AI
- ✗High-volume, repetitive decisions
- ✗Clear criteria for right/wrong
- ✗Low stakes per individual decision
- ✗Easy to review output in bulk
✓ With AI
- ✓Novel or ambiguous situations
- ✓High stakes per decision (legal, medical, financial)
- ✓Requires empathy or relationship context
- ✓Errors are costly or irreversible
Calculate your ROI
25 XPBuilding your automation habit
The biggest mistake people make isn't failing to automate — it's automating one thing and then going back to doing everything else manually. Make it a habit:
- The Friday review: Every Friday, spend 10 minutes asking: "What did I do this week that felt like a waste of my skills?" Write it down.
- The Monday build: Every Monday, spend 15 minutes automating one thing from your Friday list. Even if it's just saving a prompt template.
- The compound effect: After 3 months of this habit, you'll have 12+ automations running. That's easily 10-20 hours per month reclaimed.
Your automation action plan
50 XPBack to Marcus's Monday morning
He still opens the same five tabs. He still pastes in the same categories of data. But the part that used to take three hours — the calculations, the narrative, the deck text — now takes ten minutes.
He didn't change jobs. He didn't hire anyone. He spent 90 minutes on a Tuesday solving a problem that had been draining his best morning energy for two years. That's the whole point.
Key takeaways
- Time vampires are tasks you do repeatedly that follow the same steps. If you can explain it to a new hire in 5 minutes, AI can probably do it.
- Start with templates (Level 1), not workflows (Level 3). A saved prompt template that you paste numbers into saves 50-70% of the time and takes 15 minutes to set up.
- Batch processing is your secret weapon. Instead of running one task through AI, run fifty. AI doesn't get tired.
- Build the habit, not just the automation. The Friday review + Monday build rhythm compounds into 10-20 hours saved per month within a quarter.
Knowledge Check
1.What makes AI automation different from traditional macros?
2.Which task should you automate FIRST?
3.What is a 'time vampire' in the context of work automation?
4.Why should you include a 'human review' step in automated AI workflows?