A perfectly lit technician who never existed can now appear beside a real plumbing, HVAC, electrical, roofing, or restoration company in a matter of minutes.
That makes ad production faster. It can also create a quiet credibility problem when the image suggests a crew, vehicle, project, certification, or capability that the company cannot honestly support.
Google brought that issue closer to the customer on July 9, 2026. The company announced a new “How this ad was made” section in My Ad Center for ads across Search, YouTube, and Discover. It can indicate when an ad was created or edited with generative AI. Google will apply disclosures automatically to creative made with its own tools, while advertisers can identify AI work produced elsewhere.
I do not think contractors should respond by avoiding AI ads. I think they should get much stricter about the line between efficient production and invented proof.
AI ads for contractors now come with a provenance question
Google says the AI-label control is rolling out gradually during July across Google Ads and related advertising products. The disclosure is available globally through the ad information menu. In some places, including New York, qualifying creative may also carry a visible label directly on the ad.
The label does not automatically say an ad is dishonest. It tells a viewer that AI participated in making or editing the asset. That difference matters.
A generated background behind a real product is one thing. A generated photo that appears to document a completed electrical panel upgrade is another. A synthetic technician wearing an invented uniform beside a spotless service van may look harmless in a campaign preview, but the image can imply facts about the business.
Home-service advertising already depends on fast trust decisions. A homeowner may be comparing three companies while dealing with a failed air conditioner, a leaking water heater, or storm damage. The ad gets a few seconds to establish relevance. The landing page and business profile then have to prove that the company is real, local, capable, and available.
AI makes the first impression easier to manufacture. That raises the value of evidence after the click.
The disclosure does not repair weak or misleading creative
Google’s generated-image guidance says advertisers still need to review suggested assets for accuracy, misleading content, policy compliance, and applicable law. It also warns that an asset created with Google’s tools is not guaranteed to receive policy approval.
That is the operating detail I would pay attention to. Automation can produce an image, but it cannot confirm whether the scene reflects the company’s licenses, equipment, service area, employees, project history, or actual offer.
Google’s advertising policies prohibit misleading information and offers that are not really available. The Federal Trade Commission’s basic standard is similarly plain: advertising claims must be truthful, non-deceptive, and supported by evidence. Those standards apply to the full impression created by words and pictures, not only a headline.
I would be especially careful with four types of contractor creative:
- Project imagery: a generated kitchen, roof, mechanical room, or electrical installation that looks like completed company work.
- People and uniforms: synthetic technicians who appear to represent the actual team.
- Equipment and vehicles: invented trucks, tools, safety gear, or machinery that imply capabilities the operation may not have.
- Outcome imagery: dramatic before-and-after scenes that suggest a documented customer result.
None of this means generated assets are automatically unusable. It means the creative needs a clear role. A conceptual image can support an idea. It should not quietly impersonate evidence.
Use AI for production, then anchor the campaign in real proof
My preferred approach is to separate supporting creative from proof assets.
Supporting creative can help illustrate a service, season, homeowner problem, or general setting. Proof assets establish why this particular company deserves the call. Those include real job photos, accurate team and vehicle images, current licenses, specific service areas, verified reviews, clear warranties, honest response expectations, and a landing page that matches the advertisement.
That last part gets overlooked. A polished AI image may earn attention, but the click still needs to reach a page that continues the same promise. Our Google Ads and Local Services Ads approach treats the landing page, call tracking, service targeting, and booked-job outcome as one connected path. Creative quality matters because it starts that path, not because it replaces the rest of it.
A contractor using generated ad assets should also check whether the website provides enough real-world confirmation. If the campaign promotes trenchless sewer repair, the page should describe that service accurately, show the relevant market, explain the next step, and use real proof where proof is implied.
For an outside look at whether those website and local trust signals line up, run Revved Digital’s free 100+ point visibility audit. That can reveal gaps between the ad’s promise and the business information a homeowner finds after clicking or searching the company name.
A simple asset review before AI ads go live
I would add a short review record to every campaign using generated or meaningfully edited creative. It does not need to become a committee meeting. A five-column sheet is enough:
- where the asset came from
- whether AI created or materially edited it
- what the image and copy imply
- what evidence supports that implication
- which landing page continues the promise
This catches practical problems that a normal spelling and dimensions check will miss.
Ask whether the scene could be mistaken for company evidence
If a reasonable homeowner could interpret the image as a real employee, truck, property, or completed project, I would replace it with authentic material or make the concept unmistakable. A disclosure inside My Ad Center should not carry the entire burden of clarity.
Confirm that operations can deliver the advertised promise
An ad may say emergency response, same-day availability, free estimates, financing, or service across a wide market. The dispatcher and scheduling system need to support that claim. Generated creative can scale faster than operations, which is how a small wording choice becomes a lead-quality problem.
Measure qualified calls, not creative novelty
AI assets can make testing faster, but a higher click-through rate does not settle whether the creative helped the business. I would compare generated and authentic assets using qualified calls, booked appointments, and service fit whenever the data volume allows it.
A dramatic image may attract more curiosity while producing worse leads. A plain real-job photo may earn fewer clicks and more profitable calls. The campaign should learn from the schedule, not the art director’s favorite variation.
Real proof becomes more valuable as synthetic creative gets easier
Google’s new disclosure system does not make AI ads a bad choice for contractors. It makes provenance visible and gives advertisers a reason to create a better internal standard.
The practical takeaway is to give every generated asset one final test: does it illustrate the offer, or does it appear to prove something that never happened?
I am comfortable using AI to speed up production, explore concepts, resize assets, and support testing. I would not use it to manufacture the trust signals that should come from the business itself.
If the ad creative, landing page, service targeting, and call-quality reporting are telling different stories, book a conversation with Revved Digital about the full paid-search path. Faster creative is useful when it helps the right homeowner recognize a real company and take the next step with confidence.