Heating, Air Conditioning, and Refrigeration Mechanics and Installers
In short: today's best AI models could do about 8% of this job's working time (projected from METR's April 2026 measurement and Epoch's November 2026 index) if the work were set up for them, and about 9% by the end of 2028 if progress keeps its long-run pace. The conversations, in-person and hands-on work, about 91% of the time, stays with people.
Install or repair heating, central air conditioning, HVAC, or refrigeration systems, including oil burners, hot-air furnaces, and heating stoves.
409,670 people do this job in the US. $61,010 a year.
Reviewed on September 30, 2026: what each task is made of, and what handing it over would ask.
A typical 40-hour week
- AI could take on1 h
- A person checks AI's work2 h
- Desk work AI can't do yet0 h
- Stays with people37 h
Heating, Air Conditioning, and Refrigeration Mechanics and Installers, now: of a 40-hour week, AI could take on 1 hours, 2 hours become checking its work, 0 hours are desk work it can't do yet, and 37 hours stay with people.
The largest pieces: Estimate, fetch and install materials (about 1 h, a person checks); Study plans and specifications (under an hour, a person checks); Fill in work orders (under an hour).
Estimates for a typical week of this job, if progress keeps its long-run pace, counting work the best models finish four times in five. Every task is below.
of the working time. Now is projected from METR's April 2026 measurement and Epoch's November 2026 index.
- 4%
- 4%
- 1%
- 7%
- 83%
of the work: Anthropic's measure of what people already bring to Claude at work (2025). It counts help with part of a task, so it can be higher than the share within reach.
On the chosen pace and reliability. Estimate.
Share of the working time within reach, now to 2030
Change assumptionsHide assumptionsLong-run pace, four in five
What each choice means: , , , , . Every pace starts from .
At this date and pace, models finish tasks up to 2.4 work days long four times in five.
What handing it over would take, now
The 8% within reach splits by what a model would need. The dates, pace and reliability follow the controls above. Estimates.
of the working time. The material can be handed to a model and the result handed back.
of the working time. The model has to be connected to the system the work lives in.
is conversation short enough by length. This site leaves conversation with people, so it is not in the total.
How far this has come
With GPT-4, measured in March 2023, 0% of this job's working time was within reach at . With Claude 3.7 Sonnet, measured in February 2025, 1%. With Claude Mythos Preview (early), measured in April 2026, 6%. Today's estimate: 9%.
A backcast: today's task data with the best model METR had measured by each date. It shows how fast the models have moved, not what anyone forecast at the time.
Measured use over time
Of the Claude.ai conversations Anthropic sampled and tied to a work task, no reportable share went to this job's tasks in May 2026; 0 of its 30 tasks showed up. Within each of Anthropic's methods, never across a change: 0.013% to 0.0053% (released February and March 2025); 0.005% to 0.017% (August 2025 to February 2026).
Each release
| Conversations | Share | Rank | Tasks seen |
|---|---|---|---|
| Released February 2025 | 0.013% | 505 of 708 | 2 of 30 |
| Released March 2025 | 0.0053% | 573 of 703 | 1 of 30 |
| The week of Aug 4, 2025 (released September 2025) | 0.005% | 529 of 655 | 1 of 30 |
| The week of Nov 13, 2025 (released January 2026) | 0.012% | 452 of 672 | 2 of 30 |
| The week of Feb 5, 2026 (released March 2026) | 0.017% | 422 of 677 | 2 of 30 |
| April 2026 (released June 2026) | none | n/a | 0 of 30 |
| May 2026 (released June 2026) | none | n/a | 0 of 30 |
Measured, not estimated: Anthropic's samples of Claude.ai conversations, each tied by a model to an O*NET task, from the Anthropic Economic Index. It says where Claude's use goes, not how much of this job's working time it covers, so it is not on the same scale as the reach above. The releases differ in their sample, their classifier and the smallest share they report (the June 2026 file rounds each task to a hundredth of a percent, so a smaller task reads as none), which is why a change of method starts a new line. At least 96% of each older release's task-tied use matched this site's tasks, and all of the newest. Claude.ai only, so use through other tools and other companies' models is not in it. Every job ranked.
Robots and the physical work
Anthropic rates 80% of this job's working time as physical work. For 46% of the working time, a robot has been shown doing most of the task somewhere: 46% in a purpose-built cell. 0.46 of 3, the 324th highest of the 923 jobs.
- Not physical 20%
- No robot shown 34%
- In a purpose-built cell 46%
Anthropic's rating, not this site's estimate: Claude Opus 5, searching the web, rated each physical O*NET task by the kind of setting in which a robot it could cite (deployed, sold or demonstrated) does most of the task, weighted by its own estimate of each task's share of the time (What work can robots do?, September 30, 2026, data CC BY 4.0; every job ranked). A robot shown somewhere is not a robot in your building: cost, setup and safety decide that. The share within reach on this page still counts no body work. The robot behind each rated task is under the task below.
What changes for the people in this job
From the task-by-task review, by the end of 2028 on the long-run pace at four in five. It says which parts of the work stay with people and which change, not whether anyone keeps a job: that depends on decisions no data set can see.
What stays with people
91% of the working time: the conversation, in-person and hands-on parts of the tasks.
- Install dehumidifiers, the hands-on work, 7%
- Service HVAC systems, the hands-on work, 5%
- Repair or replace components, the hands-on work, 5%
- Connect equipment to supply lines, the hands-on work, 5%
- Wire controls to equipment, the hands-on work, 5%
- Install heat pumps, the hands-on work, 4%
What becomes checking the machine
6% of the working time: desk work a model could do by then, in tasks where a mistake costs something or a licensed person signs. A person checks it. Another 2% is desk work that is small if wrong, which can be handed over with lighter checks.
- Estimate, fetch and install materials, a licensed person signs, 3%
- Study plans and specifications, costly if wrong, 2%
- Do and recommend preventive maintenance, costly if wrong, 0.9%
What to learn
- The hands-on craft. About 83% of the time is hands-on work, which no model here does; robots are rated separately on this page.
- Signing off under your license. Some tasks need a licensed, certified or sworn person to do or sign them. The license, and the judgment it stands for, stays with the person.
- Judgment where errors are severe. Some tasks are severe if wrong. Knowing when not to trust an output matters more there than speed.
- Reviewing work you did not do yourself. The tasks under What becomes checking the machine, where a mistake costs something, stay a person's to check or sign: spotting what a draft gets wrong is a skill of its own.
Your own version of this job: read it against your job description. The headline is free; the full personal report, every task with its timeline and what to learn, is $19 (a sample).
The tasks
Sorted by . 30 tasks from .
| Task | ||||
|---|---|---|---|---|
| Estimate, order, pick up, deliver, and install materials and supplies needed to maintain equipment in good working condition.Working out what is needed, ordering it, collecting it, and fitting it.Done daily by most who do it. Robots: no robot shown. Anthropic's rating. | 7% | 1.7 h25 min to 4.2 h | Hands-on 60%, documents 20%, systems 20% | 39%Desk part half by now |
| Install dehumidifiers or related equipment for spaces that require cool, dry air to operate efficiently, such as computer rooms.Fitting dehumidifying equipment by hand.Done more than yearly by most who do it. Robots: no robot shown. Anthropic's rating. | 7% | 3.3 h1.7 h to 6.7 h | Hands-on 100% | Stays with people |
| Repair or service heating, ventilating, and air conditioning (HVAC) systems to improve efficiency, such as by changing filters, cleaning ducts, and refilling non-toxic refrigerants.Hands-on servicing: filters, ducts and refrigerant.Done daily by most who do it. Robots: no robot shown. Anthropic's rating. | 5% | 1.3 h38 min to 2.5 h | Hands-on 100% | Stays with people |
| Repair or replace defective equipment, components, or wiring.Hands-on repair and replacement.Done daily by most who do it. Robots: no robot shown. Anthropic's rating. | 5% | 1.7 h25 min to 4.2 h | Hands-on 100% | Stays with people |
| Connect heating or air conditioning equipment to fuel, water, or refrigerant source to form complete circuit.Hooking up gas, water or refrigerant by hand.Done more than yearly by most who do it. Robots: in a purpose-built cell, UltraFlex Dragon 15/20 induction brazing robot (UR-5 cobot + UltraHeat W induction supply) (sold). Anthropic's rating. | 5% | 1.7 h38 min to 3.3 h | Hands-on 100% | Stays with people |
| Lay out and connect electrical wiring between controls and equipment, according to wiring diagrams, using electrician's hand tools.Running and connecting the wiring to the diagram.Done daily by most who do it. Robots: in a purpose-built cell, Averex robotic wiring centre (Kiesling Maschinentechnik) (sold). Anthropic's rating. | 5% | 1.3 h38 min to 2.5 h | Hands-on 100% | Stays with people |
| Install or repair self-contained ground source heat pumps or hybrid ground or air source heat pumps to minimize carbon-based energy consumption and reduce carbon emissions.Fitting and repairing heat pump systems by hand.Done more than weekly by most who do it. Robots: no robot shown. Anthropic's rating. | 4% | 5 h1.7 h to 1.3 work days | Hands-on 100% | Stays with people |
| Mount compressor, condenser, and other components in specified locations on frames, using hand tools and acetylene welding equipment.Setting the heavy components on their frames by hand and weld.Done more than monthly by most who do it. Robots: in a purpose-built cell, Path Robotics AF-1 Autonomous Fit-up robotic welding cell (sold). Anthropic's rating. | 4% | 2.5 h50 min to 5 h | Hands-on 100% | Stays with people |
| Recommend, develop, or perform preventive or general maintenance procedures, such as cleaning, power-washing, or vacuuming equipment, oiling parts, or changing filters.Cleaning, oiling and changing filters, and writing up what the system needs.Done more than weekly by most who do it. Robots: in a purpose-built cell, JNW Cleaning Robot (fully-automatic heat exchanger cleaning system) (sold). Anthropic's rating. | 4% | 1.3 h25 min to 2.5 h | Hands-on 80%, documents 20% | 20%Desk part half by now |
| Install auxiliary components to heating or cooling equipment, such as expansion or discharge valves, air ducts, pipes, blowers, dampers, flues, or stokers.Fitting the auxiliary components by hand.Done daily by most who do it. Robots: no robot shown. Anthropic's rating. | 4% | 1.7 h38 min to 3.3 h | Hands-on 100% | Stays with people |
| Inspect and test systems to verify system compliance with plans and specifications or to detect and locate malfunctions.Testing the system by hand to find faults or sign it off.Done daily by most who do it. Robots: in a purpose-built cell, INFICON robotic sniffing / I-Guide3D 3D robotic leak testing system (sold). Anthropic's rating. | 4% | 1.3 h33 min to 2.5 h | Hands-on 100% | Stays with people |
| Comply with all applicable standards, policies, or procedures, such as safety procedures or the maintenance of a clean work area.Working to the safety rules and keeping the area clean through the job.Done daily by most who do it. Robots: no robot shown. Anthropic's rating. | 4% | 17 min8 min to 38 min | Hands-on 100% | Stays with people |
Within reach, per task: the share of the task's time whose instances are short enough for a model to finish at the chosen date, pace and reliability. Half by: the date half of that time comes within reach. Made of: how the task's time splits among five kinds of doing. Only the documents and systems parts can count as within reach, so a task that is 30% documents can be at most 30% within reach.
Jobs that share skills with this one
From O*NET's related occupations. Within reach now and by the end of 2028, long-run pace, four in five.
- Stationary Engineers and Boiler Operators21% now, 22% by 2028
- Plumbers, Pipefitters, and Steamfitters23% now, 23% by 2028
- Control and Valve Installers and Repairers, Except Mechanical Door1% now, 1% by 2028
- Home Appliance Repairers16% now, 16% by 2028
- Boilermakers3% now, 3% by 2028
- Solar Thermal Installers and Technicians16% now, 17% by 2028
Where this job works
The industry groups that employ the most of it (BLS, May 2025).
- Building Equipment Contractors305,330 people
- Merchant Wholesalers, Durable Goods (4232, 4233, 4235, 4236, 4237, and 4239 only)12,190 people
- Fuel Dealers8,480 people
- Local Government, excluding Schools and Hospitals6,960 people
- Commercial and Industrial Machinery and Equipment (except Automotive and Electronic) Repair and Maintenance6,790 people
Where the most of these jobs are
The metro areas with the most people in it (BLS, May 2025).
Questions about this job
- How much of the work of Heating, Air Conditioning, and Refrigeration Mechanics and Installers is within reach of AI now?
- An estimated 9% of the working time is within reach of AI models now and 9% by the end of 2028, on METR's long-run pace at four in five. Within reach means a model could finish work that long at that reliability. It is not a forecast of jobs lost or of savings.
- Which tasks of Heating, Air Conditioning, and Refrigeration Mechanics and Installers come within reach first?
- Weighted by the time each takes, the tasks with the most within reach now (projected from METR's April 2026 measurement and Epoch's November 2026 index) are: Estimate, fetch and install materials (39% of the task's time now); Study plans and specifications (50% of the task's time now); and Fill in work orders (100% of the task's time now).
- Which tasks of Heating, Air Conditioning, and Refrigeration Mechanics and Installers stay with people?
- Tasks that are people work or body work are not counted as within reach at any date. The largest: Install dehumidifiers; Service HVAC systems; and Repair or replace components. Together such tasks are about 80% of the working time, and the people and body parts of the other tasks stay with people too.
- What should Heating, Air Conditioning, and Refrigeration Mechanics and Installers learn as AI takes on more of the work?
- By the job's own task-by-task review, the first things to learn are the hands-on craft; signing off under your license; and judgment where errors are severe. By the end of 2028 on the long-run pace at four in five, about 91% of the working time stays with people, 6% becomes checking a model's work and 2% could be handed over with lighter checks. It says which parts of the work change, never whether anyone keeps a job.
- Do people already use AI for the work of Heating, Air Conditioning, and Refrigeration Mechanics and Installers?
- In the Claude.ai conversations Anthropic sampled for May 2026, no reportable share went to this job's tasks. Anthropic's labor market measure puts the share of the job's tasks seen in real Claude use at work at 2%. These are measures of use, not of how much of the work AI could finish, and they cover Claude only.
- Can robots do the physical work of Heating, Air Conditioning, and Refrigeration Mechanics and Installers?
- Anthropic's robot exposure rating (September 2026) finds 80% of the working time is physical work, and that a robot has been shown doing most of the task, somewhere, for 46% of the working time. Its index is 0.46 of 3. One example: UltraFlex Dragon 15/20 induction brazing robot (UR-5 cobot + UltraHeat W induction supply) for "connect equipment to supply lines". A robot shown somewhere is not one in every workplace; this site's own estimates still count no physical work as within reach.
- How is this estimated?
- From the 30 tasks O*NET lists for the occupation, each with an estimated length checked against GDPval and a share of the working time, reviewed one task at a time on September 30, 2026 for what each is made of (documents, business systems, conversation, in person, hands-on), and METR's measurements of how long a task the best AI models can finish, carried forward to the date from the latest, Claude Mythos Preview (early) in April 2026. Data checked November 10, 2026.
This job in your state or metro area
Embed these figures on your site
A small card with the share of working time within reach now and at the end of 2027, 2028 and 2030, the date and a link back. Free under CC BY 4.0; the card carries the credit.
Your own version of this job
The page above is the occupation. Your job is your title and what you actually do: give both, and they are read against these tasks in your browser, nothing sent anywhere. The headline is free: how much of your working time is within reach now and by the end of 2028, and how many of the tasks your description names.
The full personal report is $19, to print or save as a PDF: every task with when it comes within reach, which ones your description names, the parts of your description the task list does not cover, what stays with you, what becomes checking the machine and what to learn.