Everyone keeps saying "learn AI." And yes, being able to chat with a model is helpful, but it isn't a skill that scales. A single clever prompt won't carry you very far if the real task is doing the same thing a hundred times, consistently, with proof which usually falls to the junior exec on the team.
So here's my advice: don't stop at chatting with AI. Learn to turn AI into a system.
Picture your second week on the job as a junior marketing exec. Your manager pings you with a Google Sheet containing two hundred customer profiles. "Pick the best five for social content, I need a shortlist and reasons."
You do the obvious thing first: open ChatGPT and brainstorm content ideas. Then you stare at the sheet and feel your stomach drop. Someone has to search the company on Google, click every link, compare apples to oranges, and defend the picks.
Most teams tackle this the hard way: open the sheet and slog through 200 profiles, one by one. Or you can take the other path: treat the task like a tiny product: set clear rules, build a simple workflow, and let the system do the looping.
That's where you start to stand out.
Setting the structure
Say this project targets F&B. Your list mixes restaurants, B2B suppliers, and hawker stalls. Rather than guessing, you apply the same simple criteria to every profile:
- Business type: restaurant, hawker, or food product manufacturer?
- Cuisine: what types of cuisines do they serve?
- Online activity: active in the last 30–60 days with real engagement?
- Brand sentiment: do the recent customer reviews feel positive, negative, or meh?
A hawker stall posting fresh Malay dishes last week scores high with links to prove it. A supplier with a quiet Facebook page scores low.
Building the flow
Here's where the "bulk" part arrives. Instead of opening two hundred tabs, you wire an automation to run row by row on your Google Sheet. For each business, the flow:
- Pulls data: searches the brand, grabs the latest social post and timestamp, fetches Google Reviews, detects cuisine.
- Writes it back consistently: fills the same columns (Latest Post, Post Date, Social Link, Review Rating, Review Count, Cuisine, Evidence Links) so you're comparing like with like.
- Evaluates with an LLM: applies your small rules, suggests a one-line creative angle, marks Qualified? Yes/No, moves to the next row.
The first run will hiccup. You tweak, re-run, and end up with a clean, defensible shortlist you can trust.
From one workflow to many
Once you've built one small, repeatable workflow, you start seeing the same pattern everywhere:
- Finance: Invoice matching & reconciliation: Compare invoices to bank transactions using fuzzy rules; flag mismatches and export a clean, auditable file.
- Customer Support: Ticket triage: Tag thousands of tickets by category, urgency, and intent; auto-escalate high-risk cases.
- Operations: Supplier quote comparison: Extract price, warranty, and lead times from dozens of proposals; normalize, score, and document the pick.
- HR: Resume shortlisting with receipts: Match CVs to the JD in bulk; surface top candidates with a short "why" note and links.
- Sales: Lead research & prioritization: Run hundreds of company names through a loop to pull essentials; score with reasons and push into the CRM.
How do I start?
You don't need to be an engineer. Tools like n8n (open-source workflow builder), Zapier, and Activepieces let you chain steps together without heavy code. Google Apps Script is perfect for Sheets-driven loops.
Along the way you quietly pick up foundations that make you dangerous (in a good way): APIs, auth & rate limits, failure handling.
Sounds technical? It's time to chat with AI now. Treat AI like your on-call coach. Screenshot the UI or error, drop it into a chat, and literally ask, "What is this and what should I do next?"
My thoughts
Repetitive work drains me. Early in my career, I was lucky to explore, and I know that's harder now. Jobs are tighter, expectations are higher. That's exactly why this shift matters. Using AI in bulk doesn't just save time, it changes the work that lands on your desk. People stop asking you to "help out" and start asking you to "own" something.
Don't just chat with AI. Run it in bulk. That's how you stand out.
Originally published on LinkedIn. Want to discuss this further? Get in touch.
