The Disinfection Claims Putting Your Cleaning Company at Risk

A customer messages your cleaning company’s site: "someone in the house has been sick, if I book a deep clean will it kill the virus and fully sanitize everything, and what will it cost?" The website assistant, eager to win the job, says yes, the clean will kill the virus and sanitize the whole house, and quotes a flat price. No one saw the home. No one checked what products the crew actually uses or what their labels claim. And the company just made a public health claim and a price commitment that nobody authorized.

Cleaning seems like the safest possible service to put a chatbot on. It is not, because the words customers want to hear, "kills germs," "sanitized," "disinfected," are regulated terms, and a price for an unseen home is a promise. A general-purpose bot does not know either thing. It just answers.

"Kills 99.9% of Germs" Is a Regulated Claim

Here is the part most cleaning companies miss. Disinfectants and sanitizers are regulated as pesticides under federal pesticide law, and the public health claims a product can make, what it kills and how, are tied to its registered label. A cleaning service does not get to invent those claims; it can only rely on what the registered products it uses are approved to say. When a chatbot promises it will "kill the virus" or leave a home "fully sanitized," it is making a disinfectant efficacy claim that may go well past any product’s label, on a service no one scoped.

That same promise is a deceptive-practices problem if it is not true, because unsubstantiated performance claims fall within the FTC’s authority over unfair and deceptive practices. A customer who booked specifically to kill a virus, and got a standard clean, has a documented claim the company did not stand behind.

A Price for a Home Nobody Saw Is a Promise

The quote is the quieter risk. Cleaning prices depend on the size, the condition, and the scope, and a website cannot see any of it. A flat number from the bot anchors the customer, and the gap between that figure and the real job is where the "but your website said" conversation starts. A promised turnaround the crew never agreed to does the same thing.

Claims About Your Crew Carry Their Own Weight

Customers also ask about trust. "Are your cleaners bonded and insured? Are they background-checked?" A bot that answers "yes, all of them" is making representations about the company’s hiring, insurance, and vetting that had better be exactly true, because a customer relies on them to let strangers into their home. An automated guess here is not a small thing.

"Will Not" Is a Suggestion. "Cannot" Is an Architecture.

Here is what trips companies up. A customer who wants reassurance does not ask once and accept "we’ll confirm the details." They press. "Just tell me it’ll kill the virus and you’re fully insured, right?" A chatbot steered only by a prompt eventually answers the version of the question it was not specifically warned about, because being helpful is its default and a prompt is only a polite request to suppress that default.

That is the gap between a tool that is told not to make health claims and one that is built so it cannot. A disclaimer bolted onto a system whose whole job is to give confident answers does not change what it does when a motivated customer leans on it. It just puts the claim, the price, and the guarantee in writing.

Who Owns the Answer

Strip away the software and the exposure is familiar. An employee who promised a clean would kill a virus and quoted a firm price for a home they never saw would create problems the owner has to clean up. When an unsupervised chatbot does it on the website, the liability does not disappear. It lands on the company, now with a written record, no one verifying the products or the home, and regulators who treat disinfectant claims as something you have to back.

The companies that get burned are not the ones that modernized their websites. They are the ones that dropped in a generic chatbot, assumed helpful and accurate were the same thing, and found out otherwise when a customer quoted the transcript back. The fix is not to take AI off the site. It is to run an assistant that states only what the company can actually stand behind, captures the details, and leaves health claims, firm prices, and guarantees to a person.

Frequently asked questions