In partnership with
|
|
THE FACTORY FORWARD · 17 JUNE 2026 · ~5 MIN READ 7 rejections. 31 rejections. 22 rejections. Wrong question. Most QC teams count rejection slips. This issue shows why ranking by rupee cost changes everything your team does on Monday morning. |
|
|
|
|
|
Your quality team logged 47 rejection slips last month. Machine 3 topped the list. So they assigned it for a full maintenance check, delayed two production runs, and spent a week chasing a problem that was costing you ₹1.1L. Meanwhile, the defect with 22 slips was bleeding ₹3.2L. Nobody prioritised it. The count said it was a smaller problem. The rupee cost, so you find out three months later, said otherwise. Your QC board doesn’t care. The P&L does. This issue closes that gap. Not with a new system. With a reranking. |
|
|
ONE FIELD-NOTE The ₹18 crore plant that was solving the wrong problem A Nashik auto-component manufacturer had been running monthly QC reviews the same way for four years. The plant head would pull rejection counts from the line logs, consolidating them into a summary table by defect type. The top defect by count got a corrective action. The team moved on. One defect kept appearing at rank three or four. Low count. Nobody escalated it. It was one month late, every month, showing up just under the threshold that triggered a formal review. We ran the same three months of data through a cost-ranking prompt. Not because of AI. Because the plant head asked a different question: what is this actually costing us per unit, multiplied by volume? That rank-three defect jumped to number one. The scrap cost per unit was nearly four times the cost of the top-count defect. It had been underweighted for months because the count was low (all OD-oversize on the same part family), so it never triggered the review threshold.
|
|
ONE TOOL TO TRY The Rejection Slip Ranking Prompt You need three months of rejection data in a spreadsheet. That is the only setup required. Export your rejection log, copy the data as plain text, then paste it into Claude, ChatGPT, or Gemini with the prompt below. Step 1: Prepare your data Export your rejection log to a spreadsheet. You need at minimum: date, defect type, quantity rejected, and cost per unit (or material value). Copy the data as plain text. Step 2: Run the prompt Open Claude, ChatGPT, or Gemini. Paste the prompt below, then paste your data after the [DATA] line. The prompt does five things: GROUP by defect type, CALCULATE total ₹ cost, RANK the top 3 by cost, ROOT CAUSE HYPOTHESIS for each, and TREND direction month over month. // Rejection Slip Ranking. Paste your rejection data below You are my quality analyst. Below the [DATA] line is three months of rejection data from my manufacturing plant. Do exactly five things. Work in rupees throughout: 1. GROUP by defect type. Treat similar descriptions as the same defect (e.g. "OD oversize" and "diameter out of tolerance" are the same). List the labels you merged. 2. CALCULATE total ₹ cost per defect group. Use the cost column if present. If missing, estimate as: quantity x ₹[ask me for material cost per unit]. 3. RANK the top 3 defect groups by total ₹ cost, highest first. Show: Rank | Defect Type | Total Count | Total ₹ Cost. 4. ROOT CAUSE HYPOTHESIS: for each of the top 3, give one most likely root cause and one first thing to check on the shopfloor. 5. TREND: for each of the top 3, is this defect increasing, decreasing, or flat month-over-month? State the direction. If data is missing or unclear, tell me what you need. Do not guess. [DATA] The line doing the heavy lifting is #2 — "rank by rupee cost, not by count." Drop it and the AI re-lists your oldest invoices. Keep it, and it surfaces the defects your team has been underweighting for months.
|
|
|
|
|
|
THIS WEEK’S QUESTION When your team reviews rejection data tomorrow morning, what do you think the cost ranking would change about your current priority list? Hit reply and tell me in one line: defect type you think is underweighted right now, and why. I read every reply.
|
|
|
|
|
COMING NEXT TUESDAY You have the ranked defect list. Now what? Issue #05 covers the production planning prompt: how to turn a ranked defect list into a weekly corrective action schedule your plant head can actually run with. |
|
|
|
|
P.S. If you joined us from the June 13 AI in Manufacturing masterclass, welcome to the Factory Forward. Every Tuesday, one field-note, one tool, one number. Nothing that does not work on the shopfloor first. Until next Tuesday, Sachin Scalatris · scalatris.com |
|
|
||
|
REFER A COLLEAGUE Know a plant manager or quality head who would find this useful? Forward this issue or share your referral link below. Every referral gets you closer to tools that save real time on the shopfloor.
{{rp_personalized_text}}
|
|
|
|
|
Sachin Chiplunkar Scalatris LLP · Nashik, Maharashtra, India · scalatris.com You are receiving this because you subscribed to the Factory Forward newsletter. |
Your prompts are leaving out 80% of what you're thinking.
When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning — constraints, edge cases, examples, tone — and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.
89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Try Wispr Flow free — works on Mac, Windows, and iPhone.