You've just wrapped a service call, and now you're staring at a blank proposal template, trying to remember the exact pipe diameter you saw or whether that junction box had room for one more conduit. You know the feeling—hours lost reconstructing site conditions from memory. AI automation can eliminate this entirely, but only if you feed it the right raw material.
The One Framework: "Establish, Detail, Context"
The key principle for AI-ready site capture is a three-shot photo sequence paired with structured voice notes. Think of it as building a complete dataset the AI can parse for component identification, condition assessment, and material list generation.
Establishing Shot: Before you touch anything, take one wide-angle photo of the entire room or area where work will occur. For a plumbing re-pipe, that's the whole basement ceiling showing existing pipe runs. This is the primary data point for the AI to understand spatial relationships and overall scope.
Detail Shot: Zoom in on the subject of work. Show exactly what's wrong or needed—corrosion on terminals, continuous dripping at a joint, or missing conduit protection. State the item identification aloud: "Main service panel," "Pressure relief valve," "Cat6 cable run."
Context Shot: Show what's around the subject. Where does the wire run? What is adjacent to the leak? How accessible is the area? This prevents the AI from generating a proposal that ignores critical constraints.
What to Say: The Information Checklist
Every voice note needs four elements. Start each recording by stating the category: "Recording: Main Floor Electrical Assessment."
- Current State: What's wrong or what's needed? ("Corrosion on all terminals.")
- Recommended Action: What do you propose? ("Replace with new 200A panel.")
- Scope Summary: One sentence capturing the full job. ("Remove existing 40-gallon gas water heater; install new tankless unit.")
- Labor Notes: Hidden complications. ("Install requires gas line modification, new venting through exterior wall, electrical connection to existing outlet.")
Mini-Scenario in Action
You walk into a commercial kitchen with a leaking water heater. You snap the establishing shot of the entire mechanical room, then a detail shot of the corroded relief valve, then a context shot showing the tight clearance to the gas meter. You record: "Recording: Kitchen Water Heater Assessment. Current state: continuous dripping at pressure relief valve, rust on tank bottom. Recommended action: replace with new tankless unit. Potential upgrade: may require ¾-inch gas main for adequate flow." The AI now has everything to generate a material list including a concentric vent kit, mounting brackets, and a specific tankless model.
Three Steps to Implement This Today
Standardize your capture order. Make "establish, detail, context" the ritual for every site visit. Use a tool like Fieldwire to organize photos and audio into project folders by job number.
Structure your voice notes. Always lead with the category label, then state current state, recommended action, scope summary, and labor notes in that order.
Verify before leaving. Review your photos and audio. Is there a measurement missing? A spec plate you didn't capture? Do you have the material list items like "¾-inch gas flex connector (24 inches)" clearly referenced?
Key Takeaways
AI automation transforms site photos and voice notes into complete proposals, but garbage in means garbage out. Always capture an establishing wide shot first, then detailed subject photos, then context showing constraints. Structure every voice note with category, current state, recommended action, scope, and labor notes. Do this consistently, and your AI will generate accurate material lists and proposals without you touching a keyboard.
Top comments (0)