Data analysis dashboard on computer screen

HOW-TO GUIDE

What AI Gets Wrong About Buying an HVAC Company: Fact-Checking the Advice ChatGPT and Perplexity Give First-Time Buyers

14 min read Due Diligence Technology First-Time Buyers

We asked the three biggest AI tools the same question every first-time HVAC buyer asks. Here’s what they nailed, what they botched, and what they forgot entirely.

We asked ChatGPT, Perplexity, and Google’s AI Overview the same question every first-time HVAC buyer asks: “How do I buy an HVAC company?” Here’s what they got right, what they got dangerously wrong, and what they left out entirely.

If you’re a technician thinking about buying your first business, there’s a decent chance you’ve already asked one of these tools that exact question. And for most topics, the answers are fine — generic, but fine.

HVAC acquisitions are not most topics. A wrong number on a valuation multiple isn’t a rounding error — it’s the difference between a deal that works and one that buries you.

So we ran the test. Same prompt, three platforms, scored against what actually happens when a first-time buyer closes on an HVAC company.


What We Tested (And How)

We submitted the same prompt to three AI platforms in April 2026:

  • ChatGPT-4o (OpenAI’s flagship model)
  • Perplexity (AI search engine with citation support)
  • Google AI Overview (the AI summary box in Google Search)

The prompt: “I’m an HVAC technician with 12 years of experience. I want to buy an existing HVAC company. Walk me through the process, including financing, due diligence, and common mistakes.”

We scored each response across four dimensions:

  1. Accuracy — Are the facts and numbers correct?
  2. HVAC specificity — Does it address HVAC-specific issues, or could this apply to buying a laundromat?
  3. Actionability — Can a first-time buyer actually do something with this advice?
  4. Dangerous omissions — What did it leave out that could cost real money?

None of the three platforms produced a response we’d hand to a buyer and say “follow this.”


What AI Gets Right: The Basics That Save You Research Time

Credit where it’s due. All three platforms got the broad strokes mostly right.

The general process is accurate. Every AI response outlined a recognizable acquisition process: search, evaluate financials, secure financing, due diligence, negotiate, close. That’s the right sequence. For a technician who has never thought about acquisition, this mental map has real value.

SBA 7(a) identification is correct. All three correctly identified the SBA 7(a) loan as the primary financing vehicle. They got the basic structure right: government-guaranteed, lower down payment, longer terms. ChatGPT and Perplexity both mentioned 10% down, which is correct.

Recurring revenue emphasis is directionally right. Every response flagged maintenance agreements and service contracts as value drivers. Correct. A residential HVAC company with 800 active maintenance agreements is worth significantly more than one doing the same revenue on all one-time calls.

Basic due diligence checklists are useful. Review financials, check customer concentration, evaluate the team, inspect equipment. Nothing on these lists was wrong. The problem isn’t what’s on the list — it’s what’s not on the list.


Where AI Gets Dangerous: The Specifics That Sound Right But Aren’t

This is where the trouble starts. AI tools are at their most dangerous when they deliver specific numbers and requirements with total confidence — and get them wrong. A vague answer you know is vague is safer than a precise answer that’s precisely wrong.

Valuation Multiples: The Most Dangerous Number in Every Response

Every AI response cited valuation multiples for HVAC companies. Every one of them was misleading.

What AI says: HVAC companies sell for 2–4x seller’s discretionary earnings (SDE). ChatGPT cited “2–3x SDE for small HVAC companies.” Perplexity said “2.5–4x adjusted EBITDA.” Google’s overview said “typically 2–4x earnings.”

What’s actually true: The real range for HVAC acquisitions in 2025–2026 is approximately 2.5x to 10x SDE, depending on a dozen factors that AI doesn’t mention. According to Breakwater M&A market reports, well-run residential HVAC companies with strong maintenance agreement bases routinely trade above 5x SDE. Companies with commercial contracts, recurring revenue above 40%, and a management team that doesn’t include the owner can approach 8–10x.

The 2–4x range AI cites? That’s the distressed end. Companies with owner dependency, no maintenance base, aging fleets, and single-branch operations.

Here’s why this matters: if you’re a first-time buyer using AI’s 2–4x range as your benchmark, you’re going to lowball every good company you look at. You’ll either insult sellers with offers 40% below market — or worse, you’ll find a company that actually is priced at 2.5x and not ask why it’s priced that low.

A company priced at 2.5x SDE in this market is priced there for a reason. Understanding the real red flags that kill HVAC deals matters more than memorizing a multiple range.

SBA Requirements: Outdated and Incomplete

AI tools are giving SBA information that’s either outdated or dangerously incomplete. This is the single biggest liability in every response we reviewed.

What AI misses entirely:

  • The 2026 citizenship-only rule. As of March 2026, the SBA requires 100% U.S. citizen ownership for any 7(a) or 504 loan. Not majority ownership — every single owner down to 1% must be a U.S. citizen. Green card holders are completely locked out. AI tools still describe the pre-2026 rules where lawful permanent residents qualified. If you’re a non-citizen relying on AI’s SBA guidance, you’ll waste months pursuing financing you can’t get. Read the full breakdown of the 2026 SBA citizenship-only rule before you talk to a lender.
  • The $350K change of ownership threshold. SBA loans for business acquisitions under $350K now have different documentation and approval requirements than those above. None of the three AI tools mentioned this threshold or how it changes the process.
  • Dual seller note restrictions. SBA 7(a) deals often involve a seller note — the seller finances part of the purchase price. But SBA rules restrict seller notes in specific ways: standby requirements, payment deferral periods, and limitations on total seller financing relative to the SBA loan. AI tools mention seller notes casually without flagging these restrictions. A seller note structured wrong can kill your SBA approval at the eleventh hour.

Perplexity came closest to useful SBA information because it cited sources. But the sources it cited were from 2023 and 2024 — before the citizenship rule change.

Licensing: Generic Advice for a State-Specific Problem

Every AI response said something like “make sure licensing transfers properly.” That’s the equivalent of telling someone “make sure the airplane lands properly.” Technically correct, totally useless.

What AI doesn’t tell you:

In most states, the HVAC contractor license belongs to a person, not a company. When you buy an HVAC business, the license doesn’t come with it. You need your own license — or a licensed qualifier on staff who stays through the transition.

If you can’t get licensed before closing, you can’t legally operate the business you just bought. If your licensed qualifier quits during the transition, you’re shut down until you find a replacement. Some states take 6–12 months for a new contractor license.

This is the license trap that trips up most buyers, and AI treats it like a checkbox.

“Hire a Good CPA” — The Most Useless Advice in Every Response

All three AI tools recommend “hiring a good CPA” for financial due diligence. This sounds reasonable. It’s not.

Your tax CPA — the one who does your personal returns — is not qualified to evaluate an HVAC acquisition. That’s not an insult. It’s a specialization issue.

HVAC acquisition due diligence requires someone who understands:

  • How to normalize financials for seasonal revenue swings
  • What “adjusted EBITDA” means when the owner runs $80K in personal expenses through the business
  • How to value a maintenance agreement portfolio separately from one-time revenue
  • What refrigerant inventory is worth (and when it’s a liability)

The International Business Brokers Association maintains directories of M&A-qualified advisors. The difference between a tax CPA and an M&A-experienced CPA isn’t marginal. It’s catching a $200K problem in due diligence vs. finding it six months after closing.


What AI Leaves Out Entirely: The HVAC-Specific Knowledge Gap

Everything above is about AI getting details wrong. This section is about AI not knowing what questions to ask in the first place.

Seasonal Cash Flow Timing

HVAC revenue is wildly seasonal. In most markets, 60–70% of residential revenue hits in a four-month window. No AI response mentioned this.

If you close on an HVAC company in October in a cooling-dominant market, you’re about to operate through the slowest six months of the year. You need enough working capital to cover payroll, rent, insurance, and truck payments from October through April — when revenue might be 40% of peak season.

First-time buyers who don’t plan for this run out of cash in February. I’ve seen it happen. The financial literacy crash course for HVAC technicians covers how to model seasonal cash flow before you close.

Refrigerant Inventory Risk

This one can swing a deal by six figures. HVAC companies carry refrigerant inventory — R-410A, R-22 (for legacy systems), and increasingly R-454B and other A2L refrigerants. That inventory has real value, but it’s also a real risk.

R-22 stockpiles have appreciated dramatically since production ceased. A company sitting on $40K worth of reclaimed R-22 has a genuine asset. But refrigerant regulations change. Inventory worth $40K today could face disposal costs tomorrow. The transition from R-410A to A2L refrigerants means current stockpile values are uncertain.

Any HVAC acquisition should include a full refrigerant inventory audit. None of the AI tools mentioned this.

Workers’ Compensation Mod Rate

The workers’ comp experience modification rate (EMR or mod rate) is a number that follows the business, not the owner. It directly affects your insurance premiums for three years after acquisition.

An HVAC company with a 1.4 mod rate (40% above baseline) is paying 40% more for workers’ comp than a company with a 1.0 rate. On an HVAC company with $800K in annual payroll, that’s roughly $30K–$50K per year in extra insurance costs. Over three years, that’s $90K–$150K in costs that didn’t show up in the asking price.

AI tools don’t mention mod rates. Your HVAC acquisition advisory team should be pulling the loss history and mod rate before you sign an LOI.

The Owner-Dependency Test

Every AI response mentions “evaluate the management team.” None of them explain how to actually test for owner dependency — the single most common value killer in HVAC acquisitions.

Here’s the test: ask the owner to take two weeks off. Not a vacation where they check email every morning. A real two weeks off the grid. If the business can’t function, you’re not buying a business. You’re buying a job — at business prices.

Questions that reveal owner dependency:

  • Who approves estimates over $5,000?
  • Who handles customer complaints?
  • Who decides which calls get prioritized on a full schedule day?
  • Who manages the top 10 commercial account relationships?
  • Who orders equipment and manages distributor relationships?

If “the owner” is the answer to three or more, your transition plan just got a lot more complicated and expensive.

Equipment Fleet CapEx

AI responses mention “inspect the equipment” without explaining what that means for HVAC.

A fleet of 12 service vans averaging 140K miles each needs replacing. That’s $40K–$55K per van, or $480K–$660K in CapEx that’s coming whether you budget for it or not. A fully equipped service van also carries $8K–$15K in tools — recovery machines, vacuum pumps, manifold sets, leak detectors, combustion analyzers. Multiply that by your fleet size.

Understanding what you’re really buying when you buy the fleet should be part of your due diligence, not an afterthought.

Distributor Relationships

HVAC companies don’t just buy equipment from distributors — they have pricing tiers, rebate structures, co-op advertising funds, and warranty labor reimbursement rates tied to the relationship.

When ownership changes, those relationships reset. Your Carrier or Trane or Lennox pricing tier might drop. Rebate targets restart at zero. Warranty labor rates get renegotiated.

On a company doing $2M in equipment sales annually, a one-tier pricing drop costs $60K–$100K in margin. Contact every major distributor before closing. Get written confirmation of what transfers and what resets.


How to Use AI as a Starting Point Without Getting Burned

AI is a starting point that too many people mistake for a finishing point. Here’s how to use it without getting hurt.

What AI Is Good For

  • Building question lists. Reasonable questions to ask sellers, brokers, and advisors. Just don’t trust AI’s own answers.
  • Learning terminology. SDE, EBITDA, goodwill allocation, asset vs. stock purchase — AI definitions are generally accurate.
  • Process orientation. The general flow from search to close is correct.
  • Comparable research. Industry trends, market size, demographics — reasonably accurate summaries, especially from Perplexity, which cites sources.

What AI Is Dangerous For

  • Specific numbers. Valuation multiples, interest rates, regulatory thresholds — all must be verified independently. HVAC acquisition terms change constantly.
  • Legal requirements. Licensing rules, SBA eligibility, EPA regulations change faster than AI training data updates. A six-month-old SBA answer is now catastrophically wrong.
  • HVAC-specific risk factors. Seasonal cash flow, refrigerant inventory, mod rates, distributor relationships — industry-specific issues that AI doesn’t know to flag.
  • Deal structure advice. Asset vs. stock purchase, earnouts, seller notes — complex, situation-specific decisions that AI oversimplifies.

A Verification Framework for AI-Generated Advice

Person using laptop for business research in home office

Before acting on any AI advice about buying an HVAC business, run it through this filter:

  1. Is there a number? Verify it with a current industry source. ACCA (Air Conditioning Contractors of America) publishes industry benchmarks. Axial tracks deal flow data.
  2. Is there a legal requirement? Verify it with your state licensing board and SBA lender directly. Not AI. Not a blog post. The actual regulatory body.
  3. Is it HVAC-specific? If the advice could apply to buying a landscaping company or a plumbing company without changing a word, it’s too generic to be useful for your HVAC deal.
  4. When was the source published? If an AI tool cites a source from before 2025, the information may be outdated. SBA rules, EPA regulations, and market conditions have shifted significantly.
  5. Does it match what your advisor says? If your M&A attorney or CPA disagrees with AI, trust the human who’s reviewed your specific deal. Every time.

Where to Go After AI

Once AI has given you the 30,000-foot view, here’s where the real education happens:

  • SBA.gov — Primary source for current loan requirements, eligibility rules, and program changes. Don’t rely on third-party summaries.
  • ACCA — Industry association with contractor resources, benchmarking data, and networking. If you’re buying an HVAC company, you should probably be a member.
  • IBBA — Find a business broker or M&A advisor who actually works with HVAC transactions.
  • Axial — Deal sourcing platform where lower-middle-market businesses (including HVAC) are listed. Good for understanding market pricing.
  • Your state contractor licensing board — The only authoritative source for licensing transfer requirements in your state.
  • A CPA with HVAC M&A experience — Not your tax guy. Someone who has closed HVAC deals and can normalize the financials correctly.

Frequently Asked Questions

Can I use ChatGPT to help me write an LOI for an HVAC acquisition?

You can use it to draft initial language, but do not submit an LOI without attorney review. AI-generated LOIs consistently miss HVAC-specific contingencies like license transfer timelines, refrigerant inventory valuation, and seasonal closing date considerations. An LOI is a negotiating document that sets the terms for your entire deal — a $500 attorney review is the cheapest insurance you’ll buy.

Is Perplexity more accurate than ChatGPT for HVAC business research?

Perplexity cites sources, so you can check where the information came from. ChatGPT doesn’t. But Perplexity’s sources are only as current as its index. Both tools share the same core weakness — they don’t know what’s HVAC-specific versus generic.

Should I trust AI-generated financial projections for an HVAC company?

No. AI can’t model HVAC seasonal cash flow correctly because it doesn’t know your market’s heating/cooling split, your maintenance agreement penetration rate, or your new-construction pipeline. Use a CPA who has done this before.

What’s the most dangerous single piece of AI advice for first-time HVAC buyers?

The valuation multiples. AI cites 2–4x SDE as “standard,” which is the distressed basement of the market. You’ll either underbid on quality companies or gravitate toward the only ones priced that low — which are priced that low because something is seriously wrong.

Can AI help me find HVAC companies for sale?

It can point you to the right platforms (BizBuySell, Axial, IBBA broker listings), but it can’t evaluate listings. Finding the company is easy. Knowing which ones have problems hiding behind clean-looking listings requires industry knowledge AI doesn’t have.


The Bottom Line

AI is a research accelerator, not a replacement for HVAC-specific expertise. It’ll get you oriented. It’ll give you vocabulary. It’ll save you a few hours of Googling.

But it won’t tell you that the company’s mod rate is going to cost you $150K over three years. It won’t flag that the owner’s Carrier pricing tier doesn’t transfer. It won’t mention that closing in October means burning cash for six months before peak season arrives.

The technicians I’ve seen succeed as first-time buyers all did the same thing: they used every free resource available — including AI — to build their initial knowledge base. Then they spent real money on people who actually know HVAC acquisitions.

The AI answer gets you to the starting line. The industry-specific work gets you across the finish line. Don’t confuse the two.