You walk a job, snap a few photos of the panel, dictate a voice note about the new branch circuit, then head back to the office to spend hours piecing together a proposal. Your materials knowledge and labor pricing are second nature to you—but the AI doesn’t know them yet. The fix? Train the system on your actual data before it ever touches a customer estimate.
The Principle: Your Spreadsheet Is the AI’s Brain
The single most important step is building a structured spreadsheet that acts as the system’s lookup table. This isn’t busywork; it’s the same list you likely already keep for ordering. Use Google Sheets (or Excel) as your central repository. Set up five columns:
Column A: Item Description (e.g., “1/2” Type L Copper Pipe 10’ length”)
Column B: Your Supplier’s Item Code/SKU
Column C: Your Current Net Cost
Column D: Your Standard Selling Price (or markup percentage)
Column E: Primary Use (e.g., “Water Supply,” “Branch Circuit,” “Data Cable”)
Then define labor units for your most common tasks. For example: “Replace a GFCI outlet: 0.5 hrs, $30.” This lets the AI calculate both material cost and labor time from a single photo or note.
Putting It to Work
An electrical contractor photographs a panel that needs a new branch circuit. The AI sees the existing breakers are Eaton BR and consults your brand preference rules. It selects an Eaton BR breaker, a Halo HBU4 ceiling fan rated box, and Southwire 12/2 NM-B cable—all from your spreadsheet. Then it adds 1.5 hours of labor for running the circuit. The proposal is ready in seconds, not hours.
Three High-Level Implementation Steps
- Audit your inventory into a spreadsheet using the five columns above. Include every item you regularly use. This becomes the single source of truth for material selection.
- Create brand preference rules as simple statements (e.g., “For Cat6 data cable, always specify Belden 10GPlus” or “For all recessed LED downlights, specify Halo HLB6 unless a different trim is visible in the photo”). These rules tell the AI which brand to choose when multiple options exist.
- Define labor units for 10 repeatable tasks (e.g., “Replace water heater: 2 hrs, $180”). Then pick one past job and manually create a proposal using your new data. That proposal becomes your benchmark for validating the AI’s future outputs.
What You Gain
Your profit margins are protected because the AI applies your exact unit costs and markups every time. Errors drop dramatically—it won’t suggest a generic breaker when you always stock a specific Schneider Electric model. And you reclaim hours of back-office work, letting you focus on the jobsite and your customers. Teach the AI your trade, and it becomes your most consistent estimator.
Top comments (0)