Prompt Chain: Analyze Remakes and Draft Action Plans

What This Builds

A multi-step AI workflow that takes your raw remake data and produces: (1) a root-cause analysis identifying which dentists, case types, and time periods generate the most remakes, (2) a professional action plan for reducing your remake rate, and (3) draft talking points for conversations with dentist accounts that have recurring issues. This workflow replaces a lab manager analysis that would take 2-3 hours and turns it into a 15-minute monthly task.

Prerequisites

• Comfortable using ChatGPT or Claude for individual tasks (Level 3)
• A Google Sheet or spreadsheet with remake data (case type, dentist, date, reason)
• 15-30 cases of remake history to analyze (even just the past 3 months)
• Time needed: 1 hour to set up the chain; 15 minutes/month to run it
• Cost: Free (Claude free tier or ChatGPT free tier is sufficient)

The Concept

A prompt chain is a sequence of AI prompts where each step builds on the last one. Instead of asking one big question and hoping for a perfect answer, you break the work into steps: Step 1 organizes the data. Step 2 identifies patterns. Step 3 proposes solutions. Step 4 drafts communication.

Think of it like the way a good lab manager would think through a problem: first look at the data, then understand it, then decide what to do, then communicate the plan. The AI helps with each step in sequence.

Build It Step by Step

Part 1: Prepare Your Data

Export or manually compile your remake history into a simple format. You can use Google Sheets (if you already track this) or just type it out manually:

Remake history (last 90 days):
Case 1: Dr. Kim | PFM bridge #3-4 | Shade issue | March 2
Case 2: Dr. Kim | Crown #14 | Doesn't fit | March 8
Case 3: Dr. Rivera | Full denture | Bite adjustment | March 10
Case 4: Dr. Thompson | Crown #30 | Shade too dark | March 15
Case 5: Dr. Kim | PFM crown #2 | Open margin | March 20
[continue for all remakes]

You can use Google Sheets AI to help compile this if your data is scattered — see the Level 2 Sheets guide.

Part 2: Step 1 Prompt — Organize and Count

Open ChatGPT or Claude. Start a new conversation and paste:

I'm a dental lab manager analyzing our remake history. Here is our remake data from the last 90 days:

[paste your remake data]

Please:
1. Count the total number of remakes
2. List remakes by dentist (who has the most?)
3. List remakes by case type (crowns vs. bridges vs. dentures?)
4. List remakes by reason (shade vs. fit vs. other?)
5. Identify any patterns or concerning trends

Save the output — you'll reference it in the next step.

Part 3: Step 2 Prompt — Root Cause Analysis

In the same conversation, follow up with:

Based on this data, help me understand the likely root causes:
1. For the dentists with the most remakes — is this likely a communication/prescription issue, impression quality issue, or random variation?
2. For the case types with most remakes — is there a technical challenge specific to that case type?
3. What are the top 2-3 things our lab could do differently to reduce remakes?
4. What are the top 2-3 things we might need to ask certain dentist accounts to do differently?

Part 4: Step 3 Prompt — Action Plan

Continue in the same conversation:

Based on this analysis, draft a 1-page action plan for reducing our remake rate. Format it as:
1. Internal changes (things our lab will do)
2. Communication changes (how we'll interact with dentist accounts differently)
3. Monitoring plan (how we'll know if it's working)

Keep it practical and specific — not generic advice.

Part 5: Step 4 Prompt — Draft Dentist Talking Points

If your analysis identified a specific dentist with a high remake rate, add:

I need to have a professional conversation with [Dr. Kim's] office about their high remake rate (3 remakes in 90 days — all related to [issue]). Help me:
1. Draft 3-4 talking points for a phone call with their office manager
2. Write a brief email I could send before the call to frame the conversation constructively
3. Suggest what specific change I should ask them to make (without being accusatory)

Real Example: End-to-End Walkthrough

Data: 8 remakes in 90 days. Dr. Kim: 3 (all shade-related). Dr. Rivera: 1 (denture bite). Other dentists: 4 (mixed reasons).

Step 1 output: "Dr. Kim accounts for 37.5% of remakes, all shade-related. Most common overall reason: shade (4/8)."

Step 2 output: "Dr. Kim's consistent shade issues suggest either: (a) her office is not providing accurate shade documentation, or (b) there's a lighting mismatch between your lab and her operatory. Root cause is likely the shade prescription, not the fabrication."

Step 3 output:

• Internal: Implement shade confirmation callback for all Dr. Kim cases before starting
• Communication: Add a shade documentation form to our prescription sheets
• Monitoring: Track shade-related remakes monthly

Step 4 output: Email draft: "Dr. Kim's office, thank you for your continued partnership. I'd love to schedule a brief call to align on our shade documentation process — we want to make sure we're getting perfect results for your cases every time. Would [date/time] work? — [Lab Name]"

Time saved: 2-3 hours of lab manager analysis → 15 minutes of copy-paste and review.

What to Do When It Breaks

• AI analysis seems wrong or generic → Your data may be too sparse. Add more context: "The impression technique Dr. Kim's office uses is [type]" or "Our lab lighting is fluorescent." More context = better analysis.

• Action plan feels vague → Ask for specifics: "Make each action item concrete — what exactly will we do, by when, and who's responsible?"

• Dentist talking points feel too blunt → Ask: "Make the talking points less confrontational — frame it as improving the partnership, not pointing out their problem."

• Data is messy and hard to organize → Use Google Sheets AI first (Level 2) to clean and summarize the data, then paste the summary into ChatGPT for the analysis.

Variations

• Simpler version: Skip the prompt chain and just ask: "Here's my remake data. What's causing it and what should I do?" — you'll get a less structured answer but it's still useful.

• Extended version: Add a Step 5 that drafts a monthly lab performance report you can share with the lab owner: "Based on this analysis, draft a one-page monthly performance summary I can share with our lab director, covering remake rate, top issues, and progress against our action plan."

What to Do Next

• This week: Run this chain on your last 30 days of remake data, even if it's only 3-5 cases.
• This month: Run it monthly. The value compounds — you'll start seeing patterns over 3-6 months that are invisible in a single month.
• Advanced: Automate the data collection step using Google Sheets AI (Level 2) to auto-summarize new remake entries, then feed that summary into the prompt chain without manual compilation.

Advanced guide for dental lab technician professionals. Prompt chains work in any free AI chatbot — no paid subscription required for this workflow.