Agentforce Welcome Message: Customize Your AI Chatbot's Greeting

The welcome message is the part of an Agentforce chatbot that every user sees and almost nobody edits. Agentforce lets you build chatbots on Salesforce's AI platform for customer service and internal support, and it ships a default greeting that works and says nothing. A better one tells the user what the bot is for, what it can and cannot do, and what to type next.

This covers where the message lives, how to make it dynamic with Flow or Apex, and what actually belongs in it.

How Agentforce handles the greeting

Agentforce runs agents for everything from customer support to internal process automation. Underneath, it uses Large Language Models (LLMs) and Retrieval Augmented Generation (RAG) to produce responses that are accurate and relevant to the context they are given.

The welcome message is the first thing a user sees when they open an Agentforce chatbot, and it has four jobs:

• Identify the chatbot and say what it is for.
• Tell the user what it can and cannot do.
• Suggest a first action or question.
• Carry your brand's personality and tone.

Agentforce ships default welcome messages, and they are generic, which is the whole reason to replace them. A custom greeting lets you match the bot's persona to your organization's brand and to the kind of support you actually offer.

Where the welcome message is configured

Most of the work happens in the Einstein Bot Builder, or in whatever has succeeded it in the wider Agentforce framework by the time you read this, since it moves with the release and your configuration. That visual interface is where admins and developers configure bot flows, responses and initial greetings.

1. Configuring it in Bot Builder

If you're on Einstein Bots, the welcome message usually sits in the bot's own settings:

1. Open the Einstein Bot Builder from Setup in Salesforce.
2. Pick the bot you want to configure.
3. In the bot's configuration, look for 'Greetings', 'Welcome Message' or 'Initial Prompt'. The exact wording varies with your Salesforce edition and which Einstein Bot features are enabled.
4. Type your message into the text editor you find there. Plain text always works, and some configurations also accept basic rich text.

Say you're building a support bot for an e-commerce company and you want the greeting to earn its place.

"Hello, how can I help you?" is functional and tells the user nothing about the bot they are talking to. This does more:

Welcome to Acme Corp Support! I'm your AI assistant, ready to help with your orders, returns, or product inquiries. Please tell me what you need assistance with, or type 'menu' to see common options.

That version names the three things the bot handles, so the user knows what to ask, gives them one word to type if they would rather browse, and works the brand name in along the way.

2. Dynamic greetings with Flow and Apex

When the greeting has to change per user, Salesforce Flows or Apex do the work. That matters when the message depends on who is logged in, what they did last time, or the context they arrived with.

Flow can be triggered when a chat session begins, and you can design it to build the greeting on the spot.

A typical one does three things:

• Fetches the user's name, recent order history or support case status.
• Branches on user attributes or the channel they came in through.
• Sets the bot's initial response directly.

Say you want to greet a returning customer by name and mention their most recent order.

The flow runs roughly like this:

• Trigger on chat session start.
• Get the logged-in user's ID.
• Query the Contact or Account object for the user's name and relevant details.
• Query the Order object for the most recent order associated with the user.
• Construct a welcome message like: "Hi [User Name], welcome back! I see your recent order #[Order Number] is currently being processed. How can I assist you further today?"
• Action: Set this constructed message as the initial bot response.

Apex is the option for complex logic or a call out to an external system.

Write a class that generates the welcome message, then have your bot's backend logic invoke it.

An illustrative version, not a complete bot implementation:

public class BotWelcomeMessageService {

    @InvocableMethod(label='Generate Dynamic Welcome Message' description='Generates a personalized welcome message for the bot.')
    public static List<String> getWelcomeMessage(List<Id> recordIds) {
        List<String> welcomeMessages = new List<String>();
        // Assuming recordIds contains the User's ContactId or similar
        if (recordIds != null && !recordIds.isEmpty()) {
            Id userId = UserInfo.getUserId(); // Or determine context differently
            Contact userContact = [SELECT Id, FirstName, LastName FROM Contact WHERE Id = :UserInfo.getContactId() LIMIT 1]; // Example query

            if (userContact != null) {
                String message = 'Hello, ' + userContact.FirstName + '! ';
                message += 'Welcome to our support chat. How can I help you today?';
                welcomeMessages.add(message);
            } else {
                welcomeMessages.add('Hello! Welcome to our support chat. How can I help you today?');
            }
        } else {
            welcomeMessages.add('Hello! Welcome to our support chat. How can I help you today?');
        }
        return welcomeMessages;
    }
}

Call that method from a Flow, which then passes the generated message to the Agentforce bot. You will need to adapt UserInfo.getUserId() and the queries to however you identify the user inside the chat context.

What actually belongs in a welcome message

The same rules hold whether the message comes out of Bot Builder or out of Apex:

• Get straight to the point. Users want to know what the bot can do for them, quickly.
• State the bot's capabilities and its limits. Do not over-promise.
• Offer prompts so the user has somewhere to start. Keywords and menu options both work well.
• Match your company's tone and personality, whether that is formal, friendly or professional.
• Personalize where you can. The user's name, or a reference to their recent activity, lifts engagement noticeably.
• Keep the tone welcoming and helpful. It is the first impression the bot makes.
• Plan how you will deliver the message in other languages if your user base is global.
• Watch the interactions and the feedback, and refine the message around what works.

A message that does all of that:

👋 Hello there! I'm your friendly AI assistant for GlobalTech Solutions. 

I can help you with:

*   🚀 Tracking your order
*   💡 Product information
*   🔧 Technical support queries

Please type your question or choose an option below. If you need to speak to a human agent, just type 'human'.

It uses emojis as visual anchors, lists what the bot can do, keeps the language short, and gives the user a way out to a human.

Beyond the basic greeting

Three options worth knowing once the basic greeting is in place:

1. Contextual greetings

Sometimes the message has to change with where the user is in their journey:

• A new user gets the general introduction.
• A returning customer gets a greeting that acknowledges their history.
• Someone in the middle of a problem gets something more direct and more empathetic.

That means pulling in user session data, and possibly Salesforce's Customer 360 platform, for the richer context.

2. Internationalization (i18n)

Salesforce and Agentforce generally support standard i18n practices. Translate the welcome messages and serve them from the user's locale setting or language preference.

3. A/B testing

Show variations to different user segments and measure which one produces higher engagement, better task completion rates, or higher customer satisfaction.

What to remember

• The welcome message is the chatbot's first impression and the main place you steer the user.
• Basic customization lives in the Einstein Bot Builder interface.
• Flows and Apex handle dynamic, personalized greetings.
• Be clear, be short, and say what the bot can and cannot do.
• Match your brand's voice, and plan for internationalization if your audience is global.
• Test and refine the message instead of setting it once and walking away.