Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for energy engineers

HVAC Retro-Commissioning Fault Finder Agent

A ranked and verified list of building faults with estimated savings and checks for the site team

AnalysisScience and researchOperations
See how it worksGet access
HVAC Retro-Commissioning Fault Finder Agent: what goes in, what the agent does and what you get

What it does

Buildings waste energy through faults that show up in trend data but not in anyone's day: simultaneous heating and cooling, stuck dampers, valves leaking through. This agent reads trend logs from the building system and tests a set of fault rules. It ranks each fault by estimated energy and cost impact. Before accepting a fault, it checks whether a sensor error, such as a drifting temperature sensor, could explain the pattern. It proposes site checks, for example verifying a damper position. After the technician reports back, it updates the ranking and removes false alarms. The engineer approves the work list. Edge case: simultaneous heating and cooling appears only at night, so the agent checks the schedule before blaming a valve.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueApprovedYes, continueNoNo 1 STARTS WHEN New trend data loaded 2 USES A TOOL Read trend data and equipment schedules 3 DOES Run fault rules such as simultaneous heating andcooling and stuck dampers 4 DOES Estimate energy and cost impact for each fault 5 CHECKS THE RESULT Could sensor error or a schedule explain thepattern? If not: mark as probable false alarm and lower its rank.Back to step 3. 6 DOES Rank faults and draft site checks 7 YOU APPROVE Engineer approves the work list 8 USES A TOOL Read technician findings 9 DOES Update the ranking with confirmed and rejectedfaults 10 CHECKS THE RESULT Are the top-ranked faults confirmed on site? If not: retest the rule with better data and propose thenext check. Back to step 4. 11 RESULT Verified fault list with savings
Read the steps as a list
  1. New trend data loaded
  2. Read trend data and equipment schedules
  3. Run fault rules such as simultaneous heating and cooling and stuck dampers
  4. Estimate energy and cost impact for each fault
  5. Could sensor error or a schedule explain the pattern?If not: mark as probable false alarm and lower its rank. Back to step 3.
  6. Rank faults and draft site checks
  7. Engineer approves the work listThe agent waits here for your OK.
  8. Read technician findings
  9. Update the ranking with confirmed and rejected faults
  10. Are the top-ranked faults confirmed on site?If not: retest the rule with better data and propose the next check. Back to step 4.
  11. Verified fault list with savings

How it decides

Faults are ranked by annual cost. A fault is kept only if sensor error and schedule explanations are ruled out.

  • Rank by estimated annual cost
  • Lower rank when sensor drift above 2 degrees could explain it
  • Treat faults outside occupied hours as schedule issues first
  • Group faults on the same equipment into one work item

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Sensor error tolerance (default 2 degrees)
  • Fault rules to run
  • Energy rates
  • Minimum savings to list (default $500 a year)

What keeps you in control

It always asks you first

  • Technician work list
  • Any change to control settings

Hard limits

  • Never changes control settings
  • Never states a fault as confirmed without a site check

It stops when

  • Done: faults confirmed or dismissed
  • Stop: trend data lacks key points

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensTrend data from an office tower showed air handler 3 heating and cooling together on 41 nights, worth about 6,200 dollars a year. The agent saw the supply temperature sensor read 3 degrees off, so the false alarm check failed and it lowered the rank. A technician checked, confirmed the sensor fault and a leaking valve. The engineer approved a work order for both.

More agents for energy engineers

Analysis · Science and research

Life Cycle Inventory Data Gap Agent

Close or document every material data gap and show how it affects the result.

11 steps2 checks2 jobs
Analysis · Quality control

Utility Bill and Tariff Audit Agent

Every bill is verified against the tariff and meter data, and tariff savings are quantified.

11 steps2 checks2 jobs
Decisions · Science and research

Efficiency Measure Screening Agent

A ranked, interaction-aware shortlist of measures that meets the client's payback and savings targets.

10 steps2 checks2 jobs
Analysis · Decisions

Demand Charge Peak Shaving Agent

Lower monthly peaks through changes that do not disturb production, verified against the next bills.

10 steps2 checks2 jobs