Quick Start Guide

Get your first bot up and running in under 15 minutes. This guide walks you through creating a simple AI-powered bot, connecting a knowledge base, and deploying it to your website.

Quick Start: Build your first bot in 15 minutes

What You Will Build

By the end of this guide, you will have a working bot that:

• Greets visitors and understands their questions using AI
• Answers questions from your uploaded documents
• Can be tested in the built-in preview panel
• Is ready to embed on your website

---

Steps Overview

1

Create a Bot

Set up a new bot project in the Operations Portal with a name and description.

2

Design Your Flow

Open the visual flow builder and lay out the conversation path your bot will follow.

3

Add an AI Response Node

Drop in an LLM Response node so your bot can understand and reply to user messages.

4

Connect a Knowledge Base

Upload documents and add a KB Search node so your bot answers from your content.

5

Test Your Bot

Use the built-in test panel to have a conversation with your bot and verify it works.

6

Deploy

Publish your bot and embed the web chat widget on your website.

---

Step 1: Create a Bot

1. Log in to the Operations Portal and navigate to Bots in the sidebar.
2. Click Create Bot in the top-right corner.
3. Fill in the details:
   • Name -- Give your bot a descriptive name (e.g., "Support Assistant").
   • Description -- A short summary of what this bot does.
   • Default Language -- The primary language your bot will respond in.
4. Click Create. You will be taken to the bot builder.

imageCreate Bot form showing name, description, and language fields

The Create Bot dialog

TIP

Choose a name that makes it easy for your team to identify the bot's purpose. You can always rename it later from the bot's settings.

Step 2: Design Your Flow

The Flow Builder is where you define how your bot behaves. Every conversation follows a flow -- a series of connected nodes that control what the bot says, how it responds, and what actions it takes.

1. In the bot builder, you will see the canvas with a Start node already placed.
2. The Start node is the entry point for every conversation. All paths begin here.
3. Click the output handle (the small circle on the right side of the Start node) and drag it to an empty area to create your first connection.

Your flow will look something like this:

Start --> LLM Response --> KB Search --> LLM Response (with context)

TIP

You can also start from a Flow Template instead of building from scratch. Click the templates icon in the left panel to browse pre-built flows for common use cases like FAQ bots, lead capture, and appointment scheduling.

imageFlow builder canvas showing Start node connected to LLM Response and KB Search nodes

A simple flow with Start, KB Search, and LLM Response nodes

Step 3: Add an AI Response Node

The LLM Response node is the core of any AI-powered bot. It sends the user's message to a large language model and returns a natural language response.

1. From the Nodes panel on the left, drag an LLM Response node onto the canvas.

2. Connect the Start node's output to the LLM Response node's input.

3. Click the LLM Response node to open its configuration panel on the right.

4. Configure the node:

   • AI Model -- Select the integration to use (e.g., Claude, GPT-4, Gemini). If you have not set one up yet, go to Settings > Integrations > AI Models first.

   • System Prompt -- Tell the AI how to behave. For example:

     You are a helpful support assistant for Acme Corp. Answer questions
     clearly and concisely. If you do not know the answer, say so honestly
     and offer to connect the user with a human agent.

   • Temperature -- Controls response creativity. Use 0.3 for factual support bots, 0.7 for more conversational experiences.

5. Click Save on the configuration panel.

imageLLM Response node config panel showing AI model selector, system prompt, and temperature slider

Configuring the LLM Response node

WARNING

You must have at least one AI model integration configured before the LLM Response node will work. Navigate to Settings > Integrations > AI Models to add your API keys for Anthropic, OpenAI, or Google Vertex AI.

Step 4: Connect a Knowledge Base

A Knowledge Base lets your bot answer questions from your own documents instead of relying solely on the AI model's training data. This dramatically improves accuracy for domain-specific questions.

Create a Knowledge Base

1. Navigate to Knowledge Bases in the sidebar.
2. Click Create Knowledge Base and give it a name (e.g., "Product FAQ").
3. Upload your documents -- supported formats include PDF, DOCX, TXT, HTML, and Markdown.
4. Wait for indexing to complete. OmniBots splits your documents into chunks, generates embeddings, and stores them for fast retrieval.

Add a KB Search Node to Your Flow

1. Go back to your bot's flow builder.

2. Drag a KB Search node from the Nodes panel onto the canvas.

3. Connect it between the Start node and the LLM Response node, so the flow becomes:

   Start --> KB Search --> LLM Response

4. Click the KB Search node and configure it:

   • Knowledge Base -- Select the knowledge base you just created.
   • Top K Results -- How many document chunks to retrieve (3-5 is a good default).
   • Similarity Threshold -- Minimum relevance score to include a result (0.7 works well for most cases).

5. In the LLM Response node, update the system prompt to reference the retrieved context:

   You are a helpful support assistant for Acme Corp. Use the following
   context from our knowledge base to answer the user's question. If the
   context does not contain the answer, say so honestly.
   
   Context: {{kb_results}}

TIP

You can also connect external document sources like SharePoint, Google Drive, S3, or Dropbox to keep your knowledge base automatically synced. Go to Knowledge Bases > External Sources to set up a sync schedule.

imageKB Search node config panel showing knowledge base selector, top K, and similarity threshold settings

Configuring the KB Search node

Step 5: Test Your Bot

Before deploying, use the built-in test panel to verify your bot works as expected.

1. Click the Test button (play icon) in the top toolbar of the flow builder.
2. The test panel opens on the right side of the screen.
3. Type a message to start a conversation -- try a question that your knowledge base can answer.
4. Verify that:
   • The bot responds naturally and acknowledges the user's question.
   • Answers reference information from your uploaded documents.
   • The tone and style match your system prompt instructions.
5. Try edge cases:
   • Ask something your knowledge base does not cover -- the bot should admit it does not know.
   • Send a greeting -- the bot should respond conversationally.
   • Type gibberish -- the bot should handle it gracefully.

WARNING

The test panel uses the same AI model integration and knowledge base as production. If your API key has rate limits or usage caps, test conversations count toward those limits.

imageTest panel showing a conversation with the bot answering a question from the knowledge base

Testing the bot in the conversation simulator

Step 6: Deploy

Once you are satisfied with testing, it is time to make your bot available to users.

Publish Your Bot

1. Click Publish in the top toolbar. This saves a snapshot of your current flow as the live version.
2. You can continue editing after publishing -- changes will not go live until you publish again.

Embed the Web Widget

1. Navigate to Deploying > Web Widget in the sidebar.

2. Copy the embed snippet. It looks like this:

   <link rel="stylesheet" href="https://omnibots-widget.web.app/widget.css">
   <script src="https://omnibots-widget.web.app/widget.iife.js"></script>
   <script>
     window.omnibot.init({
       deploymentKey: 'your-deployment-key'
     });
   </script>

3. Paste the snippet into your website's HTML, just before the closing </body> tag.

4. Open your website -- you should see the chat widget in the bottom-right corner.

imageChat widget embedded on a sample website showing the bot greeting in the bottom-right corner

The deployed chat widget on your website

Customize the Widget (Optional)

You can customize the widget's appearance to match your brand:

• Colors -- Primary color, background, text colors
• Position -- Bottom-right or bottom-left
• Welcome message -- The first message users see when they open the widget
• Avatar -- Your bot's profile picture
• Language -- Widget UI language (24 languages supported, including RTL languages)

Go to Deploying > Web Widget > Customization to configure these options.

TIP

Use the browser's developer tools to preview how the widget looks on different screen sizes. The widget is fully responsive and works on mobile, tablet, and desktop.

---

Next Steps

Your first bot is live. Here is where to go next to get the most out of OmniBots:

• Flow Builder Deep Dive -- Learn about all available node types, conditional logic, variables, and advanced routing.
• Nodes Reference -- Detailed documentation for every node type including LLM Response, Condition, API Call, Tool Call, Handoff, and more.
• Tools & API Calls -- Give your bot the ability to call external APIs, look up data, and perform actions.
• Knowledge Base Configuration -- Fine-tune RAG settings like chunk size, overlap, and embedding models for better answer quality.
• CCaaS Integration -- Set up live agent handoff with Genesys Cloud, Amazon Connect, 8x8, or Google CCAI.
• Analytics Dashboard -- Track conversation volume, containment rate, and user satisfaction once your bot starts handling real conversations.
• Users & Roles -- Invite team members and assign roles to control who can build, deploy, and monitor bots.

Have questions?

If you run into any issues while building your first bot, check the specific documentation section for the feature you are working with. Each section includes troubleshooting tips and common configuration pitfalls.