Chatbot Greeting: Examples That Get Replies

A chatbot greeting is the first thing your AI says to a website visitor: the short message that pops up next to the chat button and opens the conversation. A good one does three jobs in one or two sentences: it says who is talking (your business's AI assistant), names two things it can do right now, and invites a real question. Below are greetings for twelve kinds of business, the starter questions to go with them, and your first line when you take a chat over yourself.

Ultimo Bots is an AI support agent for your website, a chat link, Messenger, Instagram, Telegram and Slack: it greets visitors in your words, answers from your own content in their language, books appointments, captures leads, hands the chat to you when you are online, and connects to Zapier, Make or n8n for everything else.

Part of the job is done for you: the website widget and the chat link show "You are chatting with an AI assistant" before the visitor types, in eight languages, and you cannot switch it off. Utah and Maine now have laws about exactly that; more on that below.

Try your greeting in a live preview

What a chatbot greeting is, and its three jobs

A chatbot greeting, often called the welcome message, is the opening text a visitor reads before typing. It usually shows up twice: as a small pop-up next to the chat button, and as the first message in the chat, under a header with your name and often an avatar. Pop-up, first message and header should all agree on who is talking.

Every good opener does three jobs:

1. Say who is talking. "I'm the AI assistant for Harbor Physio" answers the question before the visitor asks it.
2. Say what it can do right now. Two concrete things, such as finding an appointment or answering questions about prices.
3. Invite a real question. End with something the visitor can answer in their own words.

That is your chatbot introduction message. Older advice on chatbot greetings says to open with buttons, because scripted bots could only follow a menu. An AI agent reads what the visitor types, so its greeting can be short and honest.

When you set up an agent, Ultimo Bots puts in a starting welcome message and starter questions; rewrite them with the patterns below. You edit both in the design settings of your website widget, next to a live preview.

What is a good greeting message?

A good greeting message is short enough to read at a glance, specific enough that the visitor knows what to ask, and honest about who is answering. Here is one, taken apart:

Hi, I'm the AI assistant for Rosa's Salon. I can show you open times for cuts and color or answer questions about prices. What are you looking for today?

• "I'm the AI assistant for Rosa's Salon" says who is talking.
• "show you open times" is true because a booking skill uses Wix Bookings or Cal.com.
• "questions about prices" is true because the price list is on the salon's site.
• "What are you looking for today?" is open, not a yes or no.

Rules of thumb:

• One or two short sentences. The field holds 300 characters, but our pop-up is at most 320 pixels wide, so a long message becomes a block of text.
• Verbs your agent can back up. Say "book" only if it can book; with Wix Bookings say "find a time", with a Calendly link "send you our booking link".
• Specific beats friendly. "How can I help you?" is polite and says nothing about what this chat can do.

Chatbot greeting examples by business type

Most chat greeting examples online are written for button menus or for a team at the keyboard. These twelve are for an AI agent that has read your site: each says it is an AI, offers two jobs and asks a question; the last column is what makes it true, and an app counts only once a skill in the AI Studio uses it.

Copy the pattern, not the words: chatbot greeting messages work best when they sound like your own front desk. A bot greeting message that promises what the agent cannot do is worse than none, so connect the app under app actions and add a skill that uses it first; on a Wix site, the Wix app sets up skills for Bookings, Stores and Contacts during setup.

Starter questions: the three taps that do the work

Starter questions, called predefined questions in the portal, are up to three short chips just above the typing field, each up to 40 characters. Tapping one sends it as the visitor's own message, and the agent answers it like anything typed.

Take them from your inbox, not your navigation menu. Pick the three questions you answered most often last month, phrased the way customers phrase them: "Do you take my insurance?" beats "Insurance information". Give each a different job, such as a price, an opening and a policy, and tap every one before launch.

A button-based chatbot needs its chatbot buttons because it cannot read a typed sentence, so any question not on them is a dead end. An AI agent offers starter questions as shortcuts and still answers whatever else the visitor writes. That is also the short difference between AI agents and older chatbots: one follows a script, the other reads. Because Ultimo Bots answers from your own knowledge base, the three chips are the start of the conversation, not its edge.

Chatbot welcome message, pop-up and greeting: one text, three places

Many tools keep the pop-up teaser, the welcome message and the bot's first line in separate fields. In Ultimo Bots they are one text, so it has to work in three places:

1. The pop-up. Switch on attention pop-ups and choose a delay in seconds that lets the visitor see the page first; your welcome messages then appear one after another above the chat button. Once a visitor has opened the chat or closed the pop-up, it stays away for the rest of the visit.
2. The top of the chat. When the chat opens, the same messages are the first lines, just below the AI notice.
3. The agent's memory. The AI receives your welcome messages as the first thing it said in every conversation, so when a visitor simply answers your question, it knows what it asked.

You get up to three welcome messages. The first must work alone, because the pop-up shows it first; the second and third can carry detail, such as "You can write in any language; I'll answer in yours."

When you borrow chatbot welcome message examples from other tools, check which place they were written for: "Psst, got a minute?" works as a teaser but reads oddly as the agent's opening line.

How to tell if someone is a chat bot?

Visitors ask it in so many words: "Am I talking to a real person?" Answer before they do. "I'm the AI assistant for [Business]" costs six words and removes the suspicion that makes people test the chat or leave.

It is also becoming the law in more places. Maine (10 M.R.S. §1500-DD, in effect since September 24, 2025) bars businesses from using an AI chatbot or other technology in trade with consumers in a way that may lead a reasonable consumer to believe they are dealing with a human, unless the consumer is told clearly and conspicuously that they are not; a violation counts under the state's Unfair Trade Practices Act. Utah (Utah Code sections 13-77-103 and 13-77-104, enacted by S.B. 226 and in effect since May 7, 2025) requires a business using generative AI in a consumer transaction to say it is AI and not a human when a person clearly asks. It offers a safe harbor when the AI clearly and conspicuously discloses, at the outset and throughout the interaction, that it is generative AI, not human, or an AI assistant; regulated occupations must disclose high-risk uses prominently, in writing before a written interaction. California's bot law (Business and Professions Code section 17941) already bars bots that mislead people about their artificial identity to drive a sale, and a business that discloses the bot is not liable under it. For visitors in the EU, the AI Act's rule that people must be told when they are talking to an AI (Article 50(1)) has applied since August 2, 2026. As of October 2026; practical guidance, not legal advice.

On your website, Ultimo Bots gives that notice for you. The website widget and the chat link show "You are chatting with an AI assistant" above the chat before the visitor can type, in the visitor's browser language (English, German, French, Spanish, Italian, Portuguese, Dutch or Polish, otherwise English), and you cannot switch it off. A visitor can close it once read, and it steps aside while your team has taken over, so let the greeting say it too.

Five greetings that backfire

1. A human name and a stock photo on an AI. "Hi, I'm Jessica!" disappoints the moment the answers sound like software. Fix: "I'm the AI assistant for [Business]", with your logo as the avatar.
2. "How can I help you?" and nothing else. The visitor has to guess what the chat can do. Fix: name two jobs.
3. "A person will reply right away." True only while someone is online. Fix: let the AI answer; the "Talk to a human" button appears by itself while a teammate is online.
4. A menu of eight options. A scripted-bot habit that turns a chat into a phone tree. Fix: one or two sentences and up to three starter questions.
5. Promises the agent cannot keep. Booking with no calendar connected, or order tracking when your store is not on Wix. Fix: set up the skill first, or promise less.

The greeting on Messenger, Instagram, Telegram and Slack

The same agent, with the same knowledge and skills, answers in Facebook Messenger, Instagram DMs, Telegram and Slack. Those apps show no pop-up, no welcome message and no AI notice, so the agent's first reply is your greeting.

Cover that with one behavior. In the AI Studio, choose Add behavior, then Enter manually, and paste this word for word:

If none of your replies so far has said you are an AI assistant, say so briefly in this reply. The welcome message at the start does not count, since the visitor may not have seen it. If asked whether you are human, say plainly you are an AI.

Behaviors are short rules, up to 250 characters, that apply on every channel. Your agent keeps the welcome message in mind everywhere, as if it had said it, which is why the rule tells it not to count it. Keep this one channel-neutral: the agent is not told which app a message comes from, so a rule for a single app will not work reliably.

Passing a chat to a person works in the website widget and on the chat link; on Messenger and Instagram you can pause the AI for one conversation and reply from the portal.

How do I greet someone in chat?

When a visitor wants a person, the greeting is yours to write. In Ultimo Bots, a "Talk to a human" button shows in the website widget and on the chat link while someone on your team has switched themselves online in the portal; you can rename it, for example to "Talk to our team", and visitors can also just ask. When you join, the AI stops replying and the whole conversation is on your screen. While nobody is online, the button is hidden and the AI keeps answering and can take the visitor's details. More on live chat.

Your first line has different jobs from the AI's: say your name, show that you have read the conversation, and answer the open question. A few live chat greeting examples:

• "Hi Dana, Marco from the workshop here. I've read your messages: the brake job is $180, and Thursday morning works."
• "Hello, I'm Priya, one of the physios. I see you asked about a knee appointment; let's find a time that suits you."
• "Hi, Tom from [Agency] here, a real person this time. Good question about the migration; give me a moment to check."

On every paid plan, live chat translates, so you can type in English and greet a visitor in Spanish. On Boost and Ultimo, AI-suggested replies draft that first line for you, without using message credits.

Your greeting, filled in

Fill in the blanks and paste the result into your first welcome message:

Hi, I'm the AI assistant for [Business]. I can [job 1] or [job 2]. Ask me anything about [topic].

Starter questions: [the question you hear most, up to 40 characters] / [the second] / [the third]

Behavior for every channel: If none of your replies so far has said you are an AI assistant, say so briefly in this reply. The welcome message at the start does not count, since the visitor may not have seen it. If asked whether you are human, say plainly you are an AI.

Everything above except AI-suggested replies comes with every plan. Smart costs $29 a month, or $19 a month billed yearly, and includes 500 AI responses a month; every paid plan starts with a 7-day free trial, and every new agent starts on a small one-time free allowance. Details on the pricing page.

Start free for 7 days

Then test it like a visitor:

1. Open your site on your phone. Does the first welcome message make sense alone in the pop-up?
2. Tap each starter question. Would you sign every answer?
3. Ask "Am I talking to a person?" It should answer with a plain no.
4. Message your Instagram or Messenger page. The reply should say it is an AI assistant.

If all four pass, visitors know exactly who they are talking to.

Give your AI agent its first line