💬 AI Prompt Packs for HVAC, Electrical, Fire Protection, Civil, BIM, and Project Management

Why Discipline-Specific Prompting Matters

A generic prompt ("explain duct sizing") produces a generic, textbook-level answer. A well-structured prompt that specifies the project context, asks the tool to show its reasoning, and explicitly requests the tool flag uncertainty produces output that's far more useful and, critically, far easier to check for errors. The prompts below are starting points, organized by discipline, built around a consistent structure: give context, ask for reasoning shown (not just a final answer), and explicitly request the tool identify what it's uncertain about or what should be independently verified.

HVAC

• "I'm sizing a duct run for [CFM] serving [room type], with a friction rate target of [X] in. w.c./100 ft. Walk through the sizing method step by step, show which duct sizing chart or method you're using, and flag any assumption you're making about aspect ratio or velocity limits that I should confirm."
• "Explain the difference between sensible and latent cooling load for [space type/occupancy] and what design decisions each one drives — I want to understand the reasoning, not just get a number."
• "Review this equipment selection logic for a [tonnage] system and identify any step where you're uncertain or where the answer depends on a manufacturer-specific value I should verify directly against the spec sheet."

Electrical

• "Walk through an NEC Article 220 load calculation for [occupancy type, sq ft, key loads] step by step, citing which specific table or demand factor you're applying at each step so I can verify each one against the current code."
• "I'm choosing conductor size for [load, distance, conditions]. Show the ampacity lookup, any derating factors applied, and the voltage drop check separately, so I can verify each step independently rather than just the final wire size."
• "Explain the difference between GFCI and AFCI protection requirements for [specific location type] and flag that I should verify the current location list against the code edition my jurisdiction has adopted."

Fire Protection

• "Explain the hydraulic design basis for a [hazard classification] sprinkler system and what design density/area values are typically used — flag that specific values must be verified against the current NFPA 13 edition and any local amendments."
• "Walk through the logic for selecting a clean-agent suppression system vs. a water-based system for [space type/contents], showing the key decision factors rather than just a recommendation."
• "Review this fire alarm notification appliance layout description and identify any area where audibility or visibility coverage should be independently verified against current code requirements."

Civil

• "Explain the design storm return period typically used for [facility type] stormwater design and what factors would push a designer toward a more conservative return period — I want the reasoning, not just a number to use."
• "Walk through a rational method runoff calculation for [drainage area, land use, slope], showing each coefficient and its source, so I can verify each one independently."
• "Explain the difference between bearing capacity and settlement as governing criteria for shallow foundation design on [soil type] and flag which one is more likely to govern in this scenario."

BIM

• "Explain the trade-offs between LOD 300 and LOD 350 for MEP coordination modeling at the design development phase, and what specifically changes in the model between the two levels."
• "Walk through a shared-coordinate-system setup checklist for a multi-discipline linked-model project, flagging the most common setup mistakes that cause coordination problems later."
• "Review this clash detection report summary and help me categorize which clashes are likely hard clashes requiring redesign versus soft clashes that are a clearance/access concern only."

Project Management

• "Draft a project status update for [project] covering [key points], written for a client audience — flag anywhere the draft makes an assumption about schedule or budget status that I need to confirm before sending."
• "Help me structure a risk register entry for [specific project risk] — likelihood, consequence, and mitigation — using the same scoring scale our firm uses for other project risks."
• "Review this meeting summary draft against my raw notes and flag anywhere the AI-generated summary may have added detail, softened a disagreement, or drawn a conclusion that wasn't actually stated in the notes."

A General Pattern Worth Reusing

Across every discipline, the strongest prompts share three elements: specific project context (not a generic textbook question), an explicit request to show reasoning or method rather than just a final answer, and an explicit request for the tool to flag uncertainty or what should be independently verified. That third element is easy to skip and is the one that most directly supports safe, checkable AI use rather than blind trust in the output.