Quick answer: ChatGPT advertising for the trades is real — OpenAI started testing ads in ChatGPT on February 9, 2026 — but for most local home services companies the right move right now is not buying the ad. It is building a website that AI systems, ad platforms, and customers can all trust. Advertising amplifies your website. It does not fix it.
Next step for trades owners: Get a Website Foundation Audit — strategy, speed, AEO/GEO, and accessibility scored against what AI systems can actually cite.
Over the past few days I've seen a wave of posts telling business owners they need to start advertising inside ChatGPT immediately.
My answer is: maybe.
But that isn't the conversation we should be having.
The biggest shift of 2026 isn't that ChatGPT has ads. It's that your website has quietly become the foundation underneath every marketing channel you own — including the ones that don't exist yet. That is true whether you run a five-truck HVAC company, a plumbing shop, an electrical contractor, a roofing crew, or a garage door service — the trades that ChatGPT advertising will eventually try to reach.
First, the facts about ChatGPT advertising
Let's separate the news from the noise, because most of what's circulating is wrong.
OpenAI began testing ads inside ChatGPT in the United States on February 9, 2026. Per OpenAI's Ads in ChatGPT Help Center, ads appear on the Free and Go tiers only — Plus, Pro, Business, Enterprise, and Edu accounts don't see them. Ads render below the end of a response, clearly labeled as sponsored and visually separated from the answer itself. OpenAI has been explicit that ads run on separate systems from the chat model and that advertisers cannot shape, rank, or alter what ChatGPT says.
Since then the pilot has expanded to Canada, Australia, New Zealand, the United Kingdom, and Japan, with additional markets planned. Access moved from an invite-only enterprise beta into a self-serve Ads Manager beta, and in June OpenAI added a product feed feature that lets retail advertisers upload a catalog and auto-generate ads per item. Trade reporting during the first pilot put early brand commitments in the $200,000–$250,000 range — roughly double a typical experimental ad budget.
Read that last paragraph again if you run a five-truck HVAC company.
The two live formats are a shopping product carousel and contextual text placements. The buy-in, the catalog mechanics, and the agency-holding-company pilots all point the same direction: this is currently built for retail and national brands, not for a local service business selling a $14,000 system install inside a 30-mile radius.
That will change. It always does. Local Services Ads didn't exist, then they did, then they became a line item every contractor in America argues about. ChatGPT advertising for the trades will follow some version of that curve.
But right now, if you own a trades business, "should I buy ChatGPT advertising?" is the wrong question. The right question is the one underneath it.
Your website stopped being a brochure
For twenty years, businesses treated a website like an online brochure. You built it, refreshed it every three or four years, and drove traffic to it from Google Ads, SEO, social, email, and referrals.
That world is gone.
Today your website is being evaluated — continuously, automatically, without your permission — by:
- Search engines
- Large language models and answer engines
- Ad platforms calculating your Quality Score
- AI agents completing tasks on behalf of customers
- Accessibility tooling and litigation scanners
- Browsers
- Your own customers
- And now, AI advertising systems deciding what's contextually relevant
Your website is no longer a marketing asset. It is your business's digital operating system.
Every marketing dollar you spend eventually lands there. Which means every weakness in your website becomes a tax on every marketing investment you make.
That's not a metaphor. It's arithmetic. A site that converts at 1.8% instead of 3.6% doesn't cost you a little — it doubles your customer acquisition cost across every channel simultaneously. Google Ads, LSA, Meta, email, referral, ChatGPT. All of it. At once.
Advertising amplifies your website. It does not fix it. That is why a disciplined website foundation beats another channel add-on for most trades companies in 2026 — including anyone evaluating ChatGPT advertising for the trades.
AI doesn't just read your website. It judges it.
Google's AI Overviews. ChatGPT. Gemini. Perplexity. Claude. Copilot.
These systems aren't matching keywords anymore. They're forming a judgment about whether your business is trustworthy, authoritative, helpful, and technically sound — and then deciding whether you deserve to be named in an answer.
What they're actually evaluating:
| Signal category | What it means in practice |
|---|---|
| Entity clarity | Can a machine state with confidence who you are, what you do, where you operate, and who vouches for you? |
| Site architecture | Does your structure express topical relationships, or is it a flat pile of pages? |
| Structured data | Is your business, service, location, review, and author data machine-readable? |
| Content depth | Do you cover a topic completely, or do you have one thin page per keyword? |
| Demonstrated expertise | Are there real named humans with real credentials attached to the content? |
| Consistency | Does your NAP, service list, and licensing match across your site, GBP, directories, and citations? |
| Technical performance | Can the page be rendered, parsed, and understood quickly and reliably? |
| Accessibility | Can everyone — and every machine — use it? |
Here's the part most agencies still haven't internalized: AI systems can only recommend what they can confidently understand. Ambiguity is disqualifying. If a model can't determine whether you service a given city, whether you're licensed, or whether the person who wrote your furnace guide has ever touched a furnace, you don't get cited. You don't get argued about. You get skipped.
Ranking was a competition. Being cited is a qualification. That is the core of Answer Engine Optimization (AEO) for home services.
The discipline stack: SEO → AEO → GEO → AAO
We frame this for our clients as four layers, each built on the one below it. Full definitions live in our SEO vs AEO vs GEO vs AXO vs SXO guide:
- SEO — Be findable and rankable in traditional search results.
- AEO (Answer Engine Optimization) — Be extractable. Structure content so a model can lift a clean, correct, attributable answer out of it.
- GEO (Generative Engine Optimization) — Be citable. Build the entity strength and corroboration that makes a model comfortable naming you specifically.
- AAO (Agent Answer Optimization) — Be actionable. When an autonomous agent is booking the appointment, comparing the quote, or filtering the contractor list, can it actually complete the task on your site?
Most trades companies are still fighting for position on layer one while their competitors quietly build layers two through four.
And every single one of those layers resolves back to the same asset. Your website. On Hydra OS, those layers are engineered into the platform instead of bolted on later.
Site speed is not a nice-to-have
One of the most expensive mistakes I still see: obsessing over rankings while ignoring performance.
If your site takes four, five, or six seconds to become usable, you've already lost the customer. It doesn't matter what position you hold.
Google measures this through Core Web Vitals, and the thresholds aren't vague. At the 75th percentile of real user visits (web.dev Core Web Vitals):
- LCP (Largest Contentful Paint): 2.5 seconds or less
- INP (Interaction to Next Paint): 200 milliseconds or less
- CLS (Cumulative Layout Shift): 0.1 or less
INP replaced First Input Delay in 2024, and it's a far harsher test — it measures responsiveness across the whole visit, not just the first tap. Sites that quietly passed FID are failing INP right now, on mobile, where most of your calls come from.
Test it today. Open Google PageSpeed Insights or run a Lighthouse audit in Chrome DevTools. Look at the field data, not the lab score.
And don't chase a perfect 100. Understand why you're slow. The score is a symptom. The render-blocking third-party script loading before your phone number is the disease.
Why so many WordPress sites struggle
Let me be direct, because I own an agency that builds on WordPress: WordPress is not the problem.
The problem is how most WordPress sites evolve.
Over three years a site accumulates a plugin for forms, a second one for SEO, a page builder, a theme builder fighting the page builder, four tracking scripts, a chat widget, a review widget, a popup tool, a booking embed, two CSS frameworks, and a font library nobody chose on purpose.
Every addition adds complexity. Every plugin adds conflict surface. Every third-party script adds code the browser must download, parse, execute, and maintain — usually on a $200 Android phone in a driveway with two bars of service.
Eventually the website spends more time managing itself than serving the customer. We documented the pattern in our WordPress problems guide — and we ship the alternative every week on Hydra OS.
Can WordPress be fast? Absolutely. We prove it every week. But it requires real engineering discipline, an owned component library instead of plugin sprawl, continuous performance budgets, and governance over what gets added and by whom. That is not how most sites are built, and it is not what most trades owners are paying for.
The six gates every page should pass
We don't ship a page at CI Web Group until it clears six gates. I'm publishing them because I'd rather the industry get better than keep the checklist private. Our SEO intelligence and topical-authority work sit under gate two — the 8-Pillar Topical Authority framework we use across HVAC, plumbing, electrical, roofing, and garage door sites.
| Gate | The question it answers |
|---|---|
| 1. E-E-A-T | Is there demonstrable experience, expertise, authority, and trust — attached to named humans with verifiable credentials? |
| 2. SEO (8-Pillar Topical Authority) | Does this page complete a topic cluster, or does it orphan one? |
| 3. AEO | Can a model extract a clean, correct, self-contained answer from this page? |
| 4. GEO | Is the entity corroborated strongly enough elsewhere that a model will name us by name? |
| 5. CRO | Does the page make the next action obvious, low-friction, and available in under two taps on mobile? |
| 6. WCAG 2.2 AA | Can every person and every machine perceive, operate, and understand this page? |
That sixth gate deserves its own paragraph.
Accessibility has moved from ethics-only to ethics-plus-exposure. WCAG 2.2 Level AA is the operative standard in most current settlement language and procurement requirements. The DOJ's ADA Title II web rule with its 2026–2027 compliance deadlines applies to state and local government entities — not to private contractors — but private businesses are still the primary target of Title III web accessibility litigation, and demand letters have become routine in home services. Beyond the legal exposure: an inaccessible site is usually also a semantically broken site, which means it's also an AI-illegible site. Fixing accessibility improves machine comprehension as a side effect. It's the rare compliance project that pays for itself in visibility. See our ADA compliance for the trades guide for the contractor-specific path.
That's context, not legal advice. Talk to your attorney about your specific situation.
ChatGPT advertising is just another door
Will ChatGPT advertising for the trades matter? Yes. So will Google's AI search surfaces. So will Gemini, Perplexity, agentic shopping, and three things that don't have names yet.
But they are all just new doors leading people toward your business.
The question was never whether another platform would emerge. The question is whether your website is ready when someone walks through the door.
One question to ask before you spend another dollar
Before you raise your ad budget. Before you add a channel. Before you buy the AI tool the guy on stage told you about.
Ask this:
If I doubled my website traffic tomorrow, would I convert twice as many customers — or would I simply expose twice as many weaknesses?
Most owners already know the answer. They just haven't priced it.
Because the businesses that dominate the next decade won't be the ones with the biggest ad budgets. They'll be the ones with the fastest sites, the clearest authority, the strongest entities, the most accessible experiences, and the best customer journeys.
Websites built not just for search engines — but for an AI-first world.
Technology changes. Platforms come and go. Algorithms evolve.
One thing doesn't: every marketing channel eventually leads back to your website.
Build that foundation first, and everything else you do gets more effective.
Next step for trades owners: Get a Website Foundation Audit — strategy, speed, AEO/GEO, and accessibility scored against what AI systems can actually cite.
Frequently asked questions
Should trades companies buy ChatGPT advertising in 2026?
For most local home services businesses in the trades, not yet as a primary channel. The current formats — shopping carousels and contextual text placements — and the early minimum commitments are built for retail and national brands. A better use of the same dollars right now is making your website citable by AI systems, which earns you visibility in AI answers without paying per click.
What is the difference between SEO, AEO, GEO, and AAO?
SEO makes you rankable in search results. AEO (Answer Engine Optimization) makes your content extractable so a model can pull a clean answer from it. GEO (Generative Engine Optimization) makes your business citable by name through entity strength and corroboration. AAO (Agent Answer Optimization) makes your site usable by autonomous AI agents completing tasks like booking or comparing quotes.
How fast does my website need to be?
Google's Core Web Vitals thresholds, measured at the 75th percentile of real visits: Largest Contentful Paint of 2.5 seconds or less, Interaction to Next Paint of 200 milliseconds or less, and Cumulative Layout Shift of 0.1 or less. Judge yourself on field data from real users, not a lab score.
Is WordPress bad for SEO and AI visibility?
No. WordPress is not the problem — unmanaged plugin accumulation is. A disciplined WordPress build with a performance budget, a governed component library, and clean semantic markup competes with anything. A site carrying twenty plugins, two page builders, and six tracking scripts does not.
Does accessibility affect AI visibility?
Yes, indirectly and significantly. The semantic structure that makes a page accessible to a screen reader — proper heading hierarchy, real landmarks, descriptive links, labeled controls, text alternatives — is the same structure that makes a page parsable by a language model. WCAG 2.2 AA work reliably improves machine comprehension.
Why does my website affect my Google Ads performance?
Landing page experience is a direct input into Quality Score, which affects both your ad rank and your cost per click. A slow, confusing, or mismatched landing page raises what you pay for the same position and lowers what you get from the traffic.
Related reading
- ChatGPT SEO for the trades — how to earn citations inside ChatGPT answers (the organic half of ChatGPT advertising for the trades).
- AI search visibility / AEO — the commercial path for answer-engine work.
- SEO, AEO, GEO, AXO, SXO — the full discipline stack.
- Lighthouse and Core Web Vitals — how we measure site speed.
- Hydra OS — the platform behind CI Web Group delivery.
- Website design — building the foundation first.
- Accessibility statement and ADA for the trades.
- Jennifer Bagley — author page, credentials, and book path.
Sources
- OpenAI — Ads in ChatGPT (Help Center)
- OpenAI — Testing ads in ChatGPT
- web.dev — Core Web Vitals
- W3C — WCAG 2.2
- DOJ — ADA Title II web rule
Jennifer Bagley is the Founder and CEO of CI Web Group, an AI-first digital marketing agency serving more than 800 active home services companies across HVAC, plumbing, electrical, roofing, and garage doors. She is the architect of Hydra OS, the platform behind CI Web Group's delivery and intelligence operations, and of the 8-Pillar Topical Authority and AAO (Agent Answer Optimization) frameworks. With 30+ years of speaking experience and more than 1,000 industry conference engagements, she is a Six Sigma Black Belt and Certified Project Manager whose earlier career spanned enterprise systems work at Nordstrom, Fossil, and Tommy Bahama. She is the author of the forthcoming Take the Wheel: Before AI Takes It From You. Connect on LinkedIn or visit jenniferbagley.com/book.
