How to Tell If a Chatbot Is Impersonating a Real Company
A fake support chatbot only needs a familiar logo and a plausible web address. Four checks, starting with the domain the widget actually runs on, catch most impersonation scams in under a minute.
Telling whether an AI chatbot is impersonating a company comes down to four checks you can run in under a minute, before you type anything sensitive into the chat box. A fake support chatbot doesn't need to be technically advanced to work. It just needs a familiar logo, a plausible web address, and a person who assumes any chat window on a company-branded page must belong to that company. These checks test that assumption directly: the domain the widget is actually embedded on, what happens when you ask for a human, how it handles a question you can verify yourself, and what it asks you to hand over.
Why Fake Support Chatbots Are Spreading
Chat widgets are now a default part of how real companies handle support, which is exactly why AI chatbot brand impersonation scams work as well as they do. The Better Business Bureau's 2026 AI Customer Service Scam Study found scammers using AI tools to polish grammar, spin up convincing websites, and clone brand logos and customer-service phrasing, sometimes promoted through search ads made to imitate a real company's page. Close to 85% of consumers surveyed said they're already concerned about data security when using AI tools. The same tools are behind other impersonation formats too, including video convincing enough to pass as real footage.
Regulators have built a specific tool for it. The FTC's Impersonation Rule, in effect since April 2024, makes it illegal to falsely pose as a business or misrepresent affiliation with one, with civil penalties up to $53,088 per violation. The FTC ties impersonation scams broadly to $2.95 billion in consumer losses in a single year. That helps after a scam is reported. What protects you in the moment, part of a wider set of AI risks, is checking the chat window itself first.
Check the Domain the Chat Widget Is Actually On
A logo and a display name inside a chat window are just images and text the widget was configured to show, and they tell you nothing about who controls the page. Before typing anything, look at the address bar, not the chat header. If you arrived through a link in a text, email, or social post, that link may have opened a domain with an extra word, a hyphen, a swapped letter, or the wrong extension, something like support-yourbank-help.com instead of the bank's actual domain. The FTC's guidance on imposter scams is blunt about this: verify independently, using a phone number, website, or email you already know is real, never the one supplied in the message that got you there. If something feels off, close the chat and type the company's known address from memory or a saved bookmark instead of clicking anything the chat gave you.
Ask to Be Transferred to a Human
Real support operations are built around a path to a person, even when a bot handles the first reply. The BBB's scam study puts it plainly: legitimate businesses give customers a way out, a path to reach someone. Ask directly for a human agent and see what happens. A genuine escalation usually comes with something concrete: a wait-time estimate, a ticket number, or a transfer message that changes tone. A fake chatbot has no human to hand off to, so it stalls, repeating the same canned line, claiming "all agents are busy" no matter the hour, or handing you off to a "human" who answers in the exact rhythm the bot just used.
Fake channels don't only target customers. Our piece on whether a fake support ticket can trick an AI agent covers the same deception aimed at a company's own AI tools instead of its customers.
Test It With a Question You Can Verify Yourself
One of the fastest ways to spot a fake customer-service AI is to ask about something you can check independently: a specific return window, a stated shipping fee, a policy printed on your receipt. A support bot built on a company's real policy documents will answer consistently, or say it doesn't have the information and offer to find out. A fake chatbot, built to keep the conversation moving toward your details rather than to inform you, tends to do the opposite: it answers everything with confidence, inventing a plausible-sounding number rather than admitting it doesn't know. Unwarranted confidence on a question you can already check is a stronger signal than any single wrong answer.
Watch What It Asks You to Hand Over
Legitimate customer-service chatbots do not need your full card number in a plain chat message. If a payment is genuinely required, it happens through a secure payment field or a transfer to a human on a verified line, not typed into an open chat window. The same goes for a one-time passcode: that code exists so a person other than you, including whoever is on the other end of a chat, can't act on your account. Requests for remote access to your device follow the same pattern. The FTC's guidance on tech support scams describes this as one of the clearest tells: a contact you didn't initiate asking to control your screen or install remote-access software. Legitimate companies don't reach out over chat, unprompted, to tell you there's a problem that needs that kind of access.
How to Tell If an AI Chatbot Is Impersonating a Company, at a Glance
Run through this before you type anything into a support chat you didn't seek out yourself:
Check the address bar, not the chat header. A widget can display any brand name regardless of the domain it's actually running on.
Ask for a human and see what actually happens. A real handoff comes with a wait time, a ticket number, or a queue. A fake one stalls.
Ask something you can verify yourself. Confident, invented answers to a checkable question are a stronger tell than any single mistake.
Notice what it asks for. A full card number typed into chat, a one-time passcode, or a request for remote access to your device are all things a real support bot has no reason to ask for.
None of these four checks require special tools, and none take long. For fake-company AI chatbot scams that show up outside a chat window entirely, our broader guide to spotting an AI scam covers a few more patterns.
Questions
What's the single fastest check if I only have a few seconds?
Look at the actual browser address bar, not the logo or name shown inside the chat window. A widget can display any brand regardless of the domain it's running on, and a spoofed domain is the foundation the rest of the scam depends on.
Can a legitimate company's chatbot ever ask for card details?
Some do, through a secure hosted payment field used to complete a specific purchase you initiated. What it should not do is ask you to type a full card number, CVV, or a one-time passcode into an open chat message to resolve an account issue. If a chat asks for that, stop and contact the company through a number you already know is real.
Does it matter if the chatbot's grammar and tone seem off?
It's a weak signal on its own now. AI tools have made scam chatbots as fluent as legitimate ones, part of what the BBB's 2026 scam study flagged. Treat odd phrasing as a minor flag and lean on the domain check and the verifiable-question test instead.
What should I do if I already gave information to a chatbot I now think was fake?
Contact your card issuer or bank directly, using a number from the back of your card, to flag the account and watch for unauthorized activity. Change any password you reused elsewhere, and file a report at ReportFraud.FTC.gov.
How did this land?
About the author
Senior Editor, AI & Product
Cecilia leads the Swarmz editorial desk. She has spent a decade turning complex AI and product topics into writing people actually finish, and she owns the blog's quality bar.
