Your 30-Minute Quick Start Guide
Stop retyping reports. Start automating decisions. Turn messy reports into time-saving assistants. No coding. No IT tickets. Just results.
What is BUILD? A simple, proven framework to turn any manufacturing report into an AI assistant:
- B — Business Goal: What decision are you accelerating?
- U — Upload Data: What report contains the truth?
- I — Instruct Clearly: What patterns should it find?
- L — Limit Scope: What should it not do?
- D — Deploy & Iterate: Start simple, improve weekly.
Before You Start: 5-Minute Prep
- Choose your tool: ChatGPT Plus or Claude Pro ($20/month)
- Grab your messiest report (CSV, Excel, or PDF)
- Know your pain point (what chews hours every week?)
- Set a timer—30 minutes to first result
Step-by-Step BUILD Process
Step 1: BUSINESS GOAL (2 minutes)
- “Analyze sales” → “Identify customers whose purchasing dropped 20%+ this quarter”
- “Improve purchasing” → “Flag POs that are late and will stop production”
Step 2: UPLOAD DATA (3 minutes)
- Remove SSNs/credit cards; keep names, product codes, quantities
- Include ≥ 30 days of data; CSV/Excel preferred; include headers
- Big files: start with the most recent 1,000 rows
Step 3: INSTRUCT CLEARLY (10 minutes)
Magic formula: You are a [ROLE] analyzing [REPORT TYPE]. Find [SPECIFIC PATTERNS]. For each finding, provide [DESIRED OUTPUT]. Focus only on [SCOPE].
Step 4: LIMIT SCOPE (5 minutes)
- “Only analyze the data provided. Don’t make assumptions.”
- “If data is missing, say ‘Data not available.’”
- “Show your math and cite specific row numbers.”
- “Flag but don’t fix data quality issues.”
Step 5: DEPLOY & ITERATE (10 minutes)
- Run your first analysis; verify three findings
- Add one clarification; save as a Custom GPT/Project
- Schedule a weekly run
Copy-Paste Prompt Templates
Template 1: Sales Report Analyzer
You are an experienced Sales Manager analyzing our monthly sales transaction report.
Your task: Identify declines, flag trends, provide actions.
Criteria: 20%+ drop in frequency; 30%+ drop in AOV; no orders in 45+ days; 3+ returns in the past month.
For each finding: customer, metric triggered, last order date/value, recommended action.
Guardrails: analyze uploaded data only; cite row numbers; sort by revenue impact.
Output: Executive Summary → Critical Alerts → Watch List → Recommended Actions.
Template 2: PO Shortage Spotter
You are a Procurement Manager analyzing our PO status report.
Priorities: CRITICAL (zero on-hand + WO allocation), URGENT (past due/no receipt), WARNING (due in 7 days for allocated items), VENDOR RISK (multiple late deliveries).
For each finding: PO/Line, Part, Vendor, Days late, Qty short vs allocated, Action required.
Output: Stop Production Risks → Expedite Required → Monitor Closely.
Guardrails: calculate “days late” from today; only flag items in file; note “Verify stock level” if on-hand missing; vendor priority 1–5.
Template 3: Quote Generator
Produce quotes from customer requests and a pricing table (header, line items, summary, terms, and a 100-word email).
Guardrails: flag items missing in price list; use standard tier if unclear; USD 2 decimals; flag lead times > 4 weeks.
Template 4: Custom Analysis Builder
You are a [ROLE] analyzing [REPORT NAME]. Find [Pattern 1-3]. For each finding: state issue, quantify impact, cite rows, recommend an action & deadline. Output: Summary → Details → Actions → Data notes
Ready for Part 2?
