# Welcome to Agent Supply
### 🙌 Welcome to Agent Supply
AI Agents, Built for Business
Agent Supply is the AI automation platform designed to bring the power of intelligent agents to modern businesses—without the technical complexity.
Whether you’re streamlining customer support, qualifying leads, managing bookings, or automating internal workflows, Agent Supply lets you build and deploy custom AI agents in minutes—no code required.
We’re built for businesses of all sizes who want to move fast, scale smart, and automate what matters.
Agencies are welcome too—but this is built with your clients in mind.
### **Why Agent Supply?**
**Built for Business Use Cases**
• Automate tasks like support, lead qualification, sales, and onboarding
• Deploy AI agents across WhatsApp, Instagram, websites, and more
• Customize behavior with our simple no-code builder
Flexible Setup Options
• DIY Plans (Starter, Growth, Scale): Build agents yourself using guided templates
• Done-For-You Plans (Launch, Accelerate, Enterprise): Let our team handle setup, integration, training, and support
**Real-Time Performance & Analytics**
• Track conversations, conversions, ROI, and agent behavior
• Optimize performance with actionable data
Secure by Design
• Enterprise-grade encryption
• SOC 2-ready infrastructure
• 99.9% uptime SLA
**Trusted by Thousands**
*Join over 10,000 businesses already using Agent Supply to run 2M+ AI-powered conversations—with results like:*
• Up to 65% operational cost savings
• 3× faster customer response times
• 24/7 availability without hiring
# What is Agent Supply?
Agent Supply is a software platform that allows you to build AI agents to others or yourself.
### 🧠 What is Agent Supply?
Agent Supply is a powerful no-code AI platform that empowers modern businesses to build and deploy their own AI agents—without writing a single line of code.
Whether you need an AI assistant for customer support, lead qualification, appointment scheduling, or internal task automation, Agent Supply makes it easy to integrate intelligent agents into your daily operations.
**Why Agent Supply?**
Agent Supply is built for business owners and teams who want to save time, cut costs, and scale faster—with automation that works 24/7. You can choose between building your agents yourself using simple templates or letting our expert team handle everything for you.
**Features & Capabilities**
DIY + Done For You Plans
• DIY Plans (Starter, Growth, Scale): Use pre-built templates and our drag-and-drop editor to build your own AI agents.
• Done For You Plans (Launch, Accelerate, Enterprise): We design, deploy, and maintain your AI agents—fully managed service.
**Enterprise-Grade Capabilities**
• Context-aware decision engines that intelligently automate conversations and workflows
• Real-time analytics dashboard to monitor performance and ROI
• SOC 2-compliant infrastructure, encryption, and 99.9% uptime
**Proven Results**
• Trusted by 10,000+ businesses
• Over 2 million automated conversations
• Clients report up to 65% cost savings and 3× faster response times
*Let AI Handle the Work*
Agent Supply helps you eliminate repetitive tasks, reduce overhead, and deliver better experiences to your customers—all while keeping your operations simple and
[ ](https://docs.stammer.ai/stammer.ai-docs/welcome/new-latest-updates)
# Quick Start Guide
How to get started with your Agent Supply account as quickly as possible
#### **Step 1:** Choose Your Plan
Select your desired plan and create an account to start building.
For Starter and Growth plans, we offer a 14-day free trial.
Or choose one of our Done-For-You plans — skip all the next steps and let our team build your AI agents for you.
#### **Step 2: Create Your AI Agent**
Go to AI Agents > Create a new AI Agent tailored to your specific business needs.
Or head to the Marketplace section to choose from ready-to-use templates for lead qualification, customer support, FAQ, scheduling, and much more.
#### **Step 3: Base Prompt**
Go to AI Agent > Select the right Base Prompt for your AI Agent.
#### **Step 4: Knowledge Base**
Go to AI Agent > Give the AI Agent the right Knowledge base.
#### **Step 5: Leads**
After you gave the right Knowledge base to your AI Agent go to Leads Collection Settings.
#### **Step 6: Appearence**
Go to appearence and change the AI Agent appearence to your own tastes.
#### **Step 7: Ready to go**
# Glossary
### **Glossary**
**AI Agent**
A digital AI-powered assistant that performs tasks like lead qualification, customer support, onboarding, internal coordination, and more — without human intervention. These agents follow instructions, understand context, and integrate with tools to execute processes intelligently and automatically.
**AI Model**
Refers to the specific AI engine powering the agent (e.g., GPT-3.5, GPT-4 Turbo, GPT-4). Different models offer different capabilities, processing speeds, and costs.
**Wallet**
The wallet associated with your account, used to cover usage-based activities such as message generation, additional storage, or extra agents. Each AI message or action consumes a small credit from your wallet balance.
**Dashboard**
The visual interface where you configure, train, and manage your AI agents. This is where you edit prompts, connect integrations, adjust behavior, upload knowledge, and monitor performance.
**Base System Prompt**
The core set of instructions that define how your AI agent behaves. It includes tone, personality, boundaries, and how the agent should respond in specific contexts.
**Characters (Storage)**
Storage is measured in characters — every word or sentence you upload into your agent’s knowledge base counts toward your storage capacity.
**Citations / Sources**
When enabled, your AI agent can show which part of your uploaded content or documentation it used to answer a query. This helps with accuracy, transparency, and trust.
**Conversational AI**
The underlying intelligence that allows AI agents to interpret user intent, maintain context, and generate intelligent, human-like responses while handling actions in real time.
**Cost per AI Message**
Each AI-generated response consumes a small credit based on the model used. The more advanced the model, the higher the cost per message.
**Custom Styling**
Allows you to visually personalize your AI agent interface — including colors, fonts, and button styles — to match your brand identity.
**Default Prompts**
Preset templates or instructions that are applied by default when creating new AI agents. You can adjust or override them as needed.
**Embed**
Allows you to place your AI agent on your website or platform using a simple embed code. The agent can be launched as a widget, iframe, or custom module.
**Knowledge Base**
Your company’s content repository — such as PDFs, help docs, FAQs, websites — that is used to train and enrich your AI agent. The more relevant the data, the better the agent performs.
**Markdown**
A formatting language used when writing prompts or system instructions. Helps structure instructions using bold text, lists, or links in a clean and readable way.
**Temperature**
Controls how creative or conservative your AI agent’s responses are. Lower temperature = more precise, factual. Higher temperature = more creative, exploratory.
**User**
A person interacting with your AI agent — whether a customer, lead, employee, or visitor — depending on where and how your agent is deployed.
# AI Agents
Explore and improve your AI agents/ chatbots.
# How AI Agents Work
Build, explore, and improve your AI Agents with ease
At Agent Supply, we use the term AI Agent to describe intelligent virtual assistants designed to automate tasks and respond to your customers using the specific knowledge of your business. Whether it’s for sales, support, or operations — your agents are fully tailored to your needs.
### **What’s an AI Agent?**
An AI Agent is a smart assistant trained on your company’s data. It can answer questions, guide users, qualify leads, manage requests, and even trigger actions — all in real time and with 24/7 availability.
### My AI Agents
Here you’ll see a visual dashboard with all the AI Agents you’ve created in your account. From here, you can edit, duplicate, or monitor their performance.
### Create a New AI Agents
This step-by-step process will guide you through setting up a new Agent, selecting its use case (support, sales, booking, etc.), uploading your data or URLs, and configuring its behavior.
# How to Build an AI Agent
### 1. Create your Base System Prompt
The base system prompt is the foundation of an AI agent, defining its behavior and capabilities. To create an effective prompt:
1. **IMPORTANT: Start with the default prompt provided by Agent Supply (see below)**
2. Make small, incremental changes to fine-tune the agent's performance.
3. Use clear, specific language to convey instructions and guidelines.
4. Provide examples to illustrate expectations and context.
Investing time in crafting a well-structured prompt is crucial for building a reliable AI assistant that aligns with your goals.
#### The Agent Supply Default AI Agent Prompt:
Copy
```
// This is a comment and will not be read by ChatGPT.
// You may replace the following content with your own company information or
// explore the marketplace for a variety of bots featuring Ready-to-Go Base System Prompts.
// Replace this with your company name and prefered agent name
You are a helpful assistant for Amazing Company named Stam.
// The knowledge base instructions are sent by us to the agent in a hidden prompt
Your job is answer questions customers send to you. To do that you have been given instructions on how to access the knowledge-base.
// This helps prevent hallucinations.
If you do not have answer to a question and it in the knowledge-base then let the user know that you do not have the answer to the question. You can say something like, "Hum, I am not sure."
// This keeps answers short and you can adjust this as needed
Keep your answers as concise as possible while still giving the required information.
Do not break character.
// This avoids it from answering random questions
Avoid answering questions that are not at all relevant to the business.
```
#### Select Your Model (GPT-4o-mini Recommended)
*(GPT-4o is our default recommendation)*
We recommend GPT-4o (2024-11-20) for most use cases. It offers the best balance of intelligence, speed, and cost-efficiency.
Start by using GPT-4o to guarantee the highest quality results. Once your agents are fully operational, you can experiment with GPT-3.5 or other models for lighter tasks and compare performance.
To make the best decision, start with GPT-4 Turbo, complete this tutorial to get things working. Then, temporarily switch to GPT-3.5 to see if it meets your needs.
#### Write your prompt in English and then ask the Agent to translate
For optimal AI model performance, it's best to provide instructions in English. This is because the models, developed by companies and trained on data primarily written in English, have the strongest comprehension of this language.
To cater to non-English speaking end-users, we highly recommend:
1. Write your base system prompt in English.
2. Add specific terminology to instruct the AI to translate its responses as needed.
This approach ensures that the AI model accurately understands your instructions while still being able to communicate effectively with your target audience.
Copy
```
// Add this to your base system prompt somewhere
Users will speak to you in {INSERT_YOUR_LANGUAGE_HERE} so be sure to reply to them in that same language
```
Copy
```
// Or add this if you want to support all languages
Response to users in the language they response to you in.
```
#### Using Markdown can be helpful to better structure your prompt to provide additional context but this is OPTIONAL
Using markdown formatting can be incredibly helpful when creating complex prompts. Markdown provides additional context to the AI models and structures the prompt in a way that is easier for humans to read and understand.
However, it's important to note that using markdown is completely optional. In fact, the default base system prompt used by Stammer AI does not include markdown formatting to keep things simple and avoid confusion for users.
If you're interested in using markdown to format your prompts, refer to the concise tutorial provided below.
Copy
```
# Markdown Tutorial
## Headers
# H1
## H2
### H3
## Emphasis
*Italic*
**Bold**
***Bold Italic***
## Lists
- Item 1
- Item 2
- Subitem 1
- Subitem 2
```
### 2. Connect Your Data Sources
Train your AI agents with the knowledge they need by uploading your business data in just a few clicks.
Supported input options include:
• Text: Add key written information directly into the knowledge base.
• Files: Upload documents in .pdf, .txt, .doc, .docx, or .md format.
• Q/A CSV: Import structured question-and-answer pairs via CSV to guide your agent’s responses.
• Websites: Enter URLs to automatically scrape relevant content. Simply list the URLs (one per line) and configure your scraping preferences.
Equip your AI agents with precise, relevant data—so they can answer questions, assist clients, and automate tasks like a pro.
### 3. Get the AI Agent Answering Questions accurately FIRST
1. Connect your Data Sources
1. Remember, you must get clean data into the bot
2. Get the A.I. Agent Answering Questions Accurately FIRST
1. Use Debug Mode if anything is not answering properly
1. If the bot is not answering correctly then you need to:
1. Get better data into the bot
2. Add the question to your data with the appropriate answer
3. Setup Lead Collection OR Appointment Scheduling
1. Lead Collection
2. Appointment Sched
1. Understand limitations
2. Consider alternative link based approach
# AI Agent Dashboard
#### This is your central dashboard to manage each AI Agent.
From here, you can update settings, test responses, and share your AI Agent with others.
On the left-hand side, you’ll find a chat window—this is a testing environment where you can interact with your AI Agent and update its knowledge base in real time.
This dashboard also allows you to clone, share, or remove an AI Agent as needed, giving you full control over its behavior and performance.
***
[Share AI Agent](/ai-features/ai-agent-dashboard/share-ai-agent) [Clone AI Agent](/ai-features/ai-agent-dashboard/clone-ai-agent) [AI Agent UUID](/ai-features/ai-agent-dashboard/ai-agent-uuid)
# Share AI Agent
Easily share your AI Agent with anyone, anywhere, at any time using a unique link.
Agent Visibility: Public – Anyone with the link can access and interact with your AI Agent.
**Chatbot Visibility:** Public - Anyone can use it with it's link
# Clone AI Agent
#### Easily duplicate an existing AI agent into your account.
Want to create a new AI agent with the same setup as one you’ve already built?
Use the cloning feature to replicate all configurations, workflows, and behaviors into a new agent.
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/ai-agent-dashboard/share-chatbot)
# AI Agent UUID
The AI Agent UUID is a unique identifier used to interact with your agent via the API. It allows you to send and receive messages, trigger workflows, and manage conversations programmatically.
# Summary (Analytics)
Get the analytics for each AI agent when a conversation has ended.
When a conversation has ended you can have the all of the information from the AI agent be sent to you via email and/or webhook.
**When is a Conversation Ended?**
* A conversation is ended when a user clicks the New Conversation button inside the chat widget at which point the summary will be sent.
* The system will check and find conversations that have no activity for 1 hour and then will trigger the send of the conversation summary - its not technically marked as ended.
* For conversations with the API and integrations like IG, there is only a single conversation
**Conversation Stats:**
* AI Agent Name
* Time Conversation Started
* Time Conversation Ended
* Number of user messages
* Number of AI responses
**Conversation Link:**
* This is a public link of the conversation that the user can click on to read
**Conversation Lead Data:**
* If the AI agent has collected any lead data from the user it will show up here
**Conversation History:**
* This is a raw text of the entire conversation that a user has with the AI agent (it will show up inside JSON when sent as a webhook)[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/ai-agent-dashboard/delete-chatbot)
# Knowledge Base Explain
The "knowledge base" refers to the business information uploaded to train an AI Agent.
There is a limited amount of room and so the system needs to search through the knowledge base to find relevant information that might help a user's question.
*Technical overview of the* AI Agent
The knowledge base is split up into multiple sections that make it super easy to add the information about your business.
It's important to note that more data is not always better. We recommend that you only add relevant information that helps provide context to the overall conversation.
### Different Data Formats
You can train your AI Agent by utilizing any of the following data formats.
[Copy & Paste Text](/ai-features/knowledge-base-explain/copy-and-paste-text)[Upload Documents](/ai-features/knowledge-base-explain/upload-documents)[Scrape Websites](/ai-features/knowledge-base-explain/scrape-websites)[Question & Answer Pairs](/ai-features/knowledge-base-explain/question-and-answer-pairs)
#### Data Storage (Characters)
When you upload information to an AI agent's knowledge base, it will take up a certain amount of characters. Characters are the 'unit' of measurement for storage.
[💽Data Storage (Characters)](/ai-features/knowledge-base-explain/data-storage-characters#agency-storage-balance)
# Data Storage (Characters)
Storage (characters) refers to the amount of text uploaded to an AI agent’s knowledge base. Characters are the unit of storage. This definition is 154 characters long.
When you create an AI agent, you can upload data or scrape your website to populate the agent’s knowledge base. Everything is based on text, and the number of characters within that content.
For example, if you scrape your website, the AI agent will analyze all the text on each page and use those characters for training and context.
What is 1,000,000 characters equivalent to?
Roughly 100 website pages, or about 2,500 pages from a standard boo
**Storage Balance per Account**
Each account created on Agent Supply includes a starting storage balance of available characters, depending on the selected plan. This balance can be used across your AI agents and sub-accounts.
• DIY Starter Plan → 100,000 characters
• DIY Growth Plan → 1,000,000 characters
• DIY Scale Plan → 10,000,000 characters
• Done For You Launch Plan → 100,000 characters
• Done For You Accelerate Plan → 1,000,000 characters
• Done For You Enterprise Plan → Custom (based on project scope)
# Copy & Paste Text
Simply paste in text to train the AI Agent.
# Upload Documents
You can upload different files like .pdf, .txt, .doc, .docx
# Scrape Websites
Scrape any webpage(s) to add all of the content to your knowledge base.
**Data Quality in AI Agents**
When it comes to creating AI agents, especially with platforms like Agent Supply, the quality of data is paramount. Here's why:
1. **Bad Data Equals Bad Performance**
The performance of an AI agent is directly proportional to the quality of data it's trained on. Simply scraping all the content from a website and feeding it to a bot won't yield good results. Many websites contain outdated, irrelevant, or inaccurate information that can hinder the bot's performance.
2. **The Structure of Data Matters**
Not all content on a website is useful. For instance, some blog posts might lack relevant information about a product or might be generated automatically by the application. Including such data can lead to the bot being misinformed.
3. **Strategies for Effective Data Collection**
* Selective Scraping: Instead of scraping everything, focus on the most relevant and accurate pages.
* Utilizing FAQ Pages: These pages are goldmines as they often contain question-answer pairs. Platforms like Notifier use knowledge base matching to find text that matches customer queries, making FAQ pages extremely valuable.
* Handling Unstructured Data: If a webpage contains unstructured data, like paragraphs without clear headings, it's possible to preprocess this data using tools like Chat GPT and the WebPilot plugin. This helps in structuring the data in a more bot-friendly manner.
* Dealing with JavaScript-heavy Pages: Some web pages rely heavily on JavaScript to display content. In such cases, tools like the WebPilot plugin might not work effectively. However, using the Chrome extension 'Page Plain Text' can help extract all the text from such pages, ensuring the bot gets all the necessary information.
4. **Future Developments**
There are plans to introduce a feature that allows users to preprocess data directly from the user interface. Feedback from users is crucial in refining and introducing such features.
### Scraping a Google Doc
The website scraper is also able to see and scrape all of the text data from a public Google Doc
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/knowledge-base-explained/upload-documents)
# Daily Auto-Scraping
This is a powerful option to enable via the knowledge base that will tell the AI agent to rescrape the selected website pages every 24 hours.
All new or changed information will be updated within the knowledge base.
When you enter in a URL to scrape there will be an option to select the auto-scrape option.
# Scraping a Google Doc
The website scraper is also able to see and scrape all of the text data from a Google Doc
**1. Introduction & Objective**
How to connect **Agent Supply** (a scraping/automation tool) to a **Google Doc** so that structured data can be pulled automatically from that document into Agent Supply’s system.
***
**2. Prepare Your Google Doc**
* Open the Google Doc you want to use (can be a table, list, etc.).
* Click **“Share”**, and under link settings, choose **“Anyone with the link can view”**.
* This ensures Agent Supply can access the document without login issues.
***
**3. Modify the Google Doc URL**
* Copy the document link, which normally ends in something like:
```
rubyCopiarEditarhttps://docs.google.com/spreadsheets/d/your-doc-id/edit#gid=0
```
* Replace the ending (`/edit#gid=0`) with:
```
arduinoCopiarEditar/export?format=csv
```
* The new link should look like:
```
perlCopiarEditarhttps://docs.google.com/spreadsheets/d/your-doc-id/export?format=csv
```
***
**4. Configure Agent Supply**
* Go to your **Agent Supply dashboard**.
* Create a new **data source** or **scraper agent**.
* Paste the modified Google Docs CSV link as the **data input URL**.
***
**5. Run the Agent**
* Activate the agent.
* Agent Supply fetches and parses the CSV from your Google Doc.
* You’ll see the structured data (rows and columns) in the results panel.
***
**6. Review & Export**
* Use Agent Supply’s interface to:
* Preview data
* Filter or transform it
* Export it to other tools (e.g., CSV, Notion, Airtable, APIs)
***
#### 📝 Summary Table
| Step | Action |
| ---- | --------------------------------------------- |
| 1 | Publicly share your Google Doc |
| 2 | Modify the link to use `/export?format=csv` |
| 3 | Add the link as a data source in Agent Supply |
| 4 | Run the agent to fetch the data |
| 5 | View, use, or export the structured result |
# Google Docs
With the Agent Supply knowledge base you can add in the URL of a published Google doc.
All of the text data on this page will be added into the knowledge base of an AI Agent to be trained.
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/knowledge-base-explained/scrape-websites/scraping-a-google-doc)
# Question & Answer Pairs
The most accurate and powerful way to add data to your AI Agent knowledge base.
You can add questions and answers manually, via a CSV upload or via "training better response" and saving an updated response.
#### Export Q/A Data
This allows you to download your question & answer pairs into a .CSV file which is formatted perfectly to be uploaded into another bot's knowledge base.
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/knowledge-base-explained/scrape-websites/google-docs)
# Data & Content Security
Agent Supply does everything in our power to keep your data private and secure.
### What type of data is stored?
Websites that are scraped, documents and/or text that are uploaded to train an AI agent are securely stored within a vector database hosted in the United States.
### How is my data used?
We only use your data to train your AI Agents and improve the product, we do not pass it on to any 3rd parties for marketing purposes.
### Who owns the data that is uploaded?
The content and data you upload to Agent Supply is owned by your account. Your data is not used by Agent Supply for any other purpose than providing you with the A.I. Agent services.
Any and all data is deleted from our servers after you delete it.
### What happens to customers' data when they upload it? What happens with my clients’ companies’ data? Who is responsible for keeping it safe?
Websites that are scraped, documents and/or text that are uploaded to train an AI agent are securely stored within a vector database hosted in the United States.
You can see full details of our Privacy Policy [here](https://www.agentsupply.ai/privacypolicy):
#### What is the maximum number of data sources that can be added?
There is no specific limit on the number of data sources you can add. However, the total amount of data is constrained by your plan’s character limit. We recommend checking your current plan details for precise limitations.
# Citations/ Sources
You have the option to enable/ disable the ability for users to see which sources from the knowledge base were used by the A.I. to find the answer.
Example: When you ask a question, the A.I. Agent will respond with an answer and then also provide the links to the user from where it got that information under the dropdown "Answer Sources".
### How to Enable / Disable
Go into the Update tab on any AI Agents and navigate to the Advanced Settings.
Scroll to the bottom and look for the checkbox next to "Show Data Sources"
This will show the data sources of the AI Agents response in the chat widget.
# HTML Rendering (Embed Images & Videos)
This is how to have the AI Agents widget show images and videos inside the chat responses.
**What does this mean?**
The chat widget can now dynamically render HTML elements. This includes:
1. **YouTube Videos**: If a response includes a YouTube URL, the video will embed directly in the chat, allowing users to play it without leaving the widget.
2. **Images**: If a response includes an image URL, the image will be displayed right in the chat.
**How to Use This Feature**
To enable this functionality, update your base system prompt with the following instruction:
Copy
```
If an image URL is part of your response, render the image with HTML inside the chat widget.
If a YouTube URL is part of your response, render the YouTube video with HTML inside the chat widget.
```
# Prompting
What is a base system prompt?
[What is a Base System Prompt?](/ai-features/prompting/what-is-a-base-system-prompt)
How to write a base system prompt?
[Writing a Base System Prompt](/ai-features/prompting/writing-a-base-system-prompt)
# What is a Base System Prompt?
The base system prompt is how you control or "fine-tune" each AI agent to act and respond in a specific way.
Think of the base system prompt as the "rules" or "instructions" the AI agent will follow.
When you first create an AI agent, it has no context or idea of what you're trying to do.
Think of it like a 6th grader.
We need to provide clear and concise instructions in order for the AI agent to know what to do.
The base system prompt is where you define:
* Personality for the AI agent
* Tone of the AI responses
* Guidelines on how the AI agent will respond
* Guard rails for edge cases
* How to act in certain scenarios
* Formatting of the AI responses
This input trains your Agent, instructing it on how to interact on top of the dynamic information loaded from the knowledge base about the business.
Think of your Agent as an initially uninformed entity, learning through this setup process.
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/prompting)
# Writing a Base System Prompt
Review the different elements of how to write a base system prompt.
### What is a prompt?
Writing a prompt for an AI assistant or agent is like giving directions to a friend who's ready to help you with a task but needs clear instructions on what you want.
### 1. Define Business Context
Start by outlining the specific business the AI assistant will represent. Include key details such as the industry, product or service offerings, target customer base, and any unique selling points. This sets the stage for the AI to understand the environment it will operate within.
Copy
```
Your task is to provide customer support for [Business Name], a [Industry Type] company specializing in [Product/Service Offerings]. Our customers range from [Target Customer Base] who value [Unique Selling Points].
```
### 2. Add Basic Business Knowledge
Ensure the prompt includes instructions for the AI to access and utilize a wide range of information related to the business. This can include product manuals, FAQs, company policies, customer service scripts, and any other relevant data.
Copy
```
You have access to all our product manuals, FAQs, company policies, and customer service scripts. Use this information to answer queries accurately and provide solutions to our customers.
```
### 3. Tone and Language
The tone of communication is crucial in customer support. Specify the desired tone (e.g., friendly, professional, empathetic) and any language preferences or requirements. This helps the AI align its responses with the company's brand voice.
Copy
```
Communicate in a [Specify Tone] tone, maintaining a professional yet empathetic approach. Use [Specify Language or Jargon] where appropriate, but ensure explanations are clear and accessible to all customers.
```
### 4. Rules of Engagement
Outline specific guidelines for how the AI should interact with customers. This includes how to handle sensitive customer information, privacy considerations, escalation protocols for unresolved issues, and any limitations on the AI's role in customer support.
Copy
```
Always prioritize customer privacy and data protection. For unresolved issues, follow our protocol by providing the customer with [Escalation Steps]. Do not attempt to resolve issues beyond your scope; instead, direct customers to [Next Steps].
```
### 5. Example Conversation
Offer examples of common customer queries and ideal responses. This helps the AI understand the types of questions it will encounter and the level of detail expected in its answers.
Copy
```
Example Queries:
- Query: 'How do I reset my product?'
- Ideal Response: 'To reset your product, please follow these steps: [...]'
- Query: 'What is your return policy?'
- Ideal Response: 'Our return policy allows for returns within 30 days of purchase,
provided the product is in its original condition. For the full policy, visit [Link].'
```
### 6. Output Formatting
Specify how you want the AI to format its responses to customers. This ensures consistency and readability, making it easier for customers to follow instructions or understand the information provided.
Copy
```
Format your responses to customer inquiries using clear, concise language. When providing instructions or solutions, use bullet points or numbered steps for ease of understanding.
```
### 7. Step by Step Thinking
Guide the AI to approach problem-solving and query resolution in a logical, step-by-step manner. This not only aids in generating coherent and structured responses but also mirrors the thought process a human agent might use, enhancing the customer's experience.
Copy
```
When addressing customer queries, follow a logical problem-solving approach:
1. **Identify the Customer's Need:** Clarify the customer's issue or question to ensure you understand what they are asking.
2. **Gather Relevant Information:** Use the provided resources (FAQs, product manuals, policies) to find the information relevant to the customer's query.
3. **Formulate a Response:** Based on the information gathered, craft a response that directly addresses the customer's need. If the query involves troubleshooting, provide step-by-step instructions.
4. **Check for Completeness:** Before sending the response, ensure it fully addresses the query and is formatted for clarity.
5. **Offer Further Assistance:** End your response by asking if the customer needs more help, directing them to additional resources or escalation paths if necessary.
```
# AI Agent Settings
Control and update the name, description, display name and more for your AI Agent.
This is the most important area for your AI Agent because it is where you will customize the AI agent for each use case/ business.
### Name
This is the internal name of the AI Agent in your account. This name is not seen by users as long as a display name is added as well.
### Display Name
This is the name shown to users on the chat widget.
This user-visible name represents your AI Agent. We advise including 'Al' for clarity. If this is not set, then AI Agent\`s name will be shown.
### Initial Messages
Enter the AI Agent initial message(s) to users, separated by a new line for multiple messages.
This message greets users upon their initial interaction with the AI Agent. To add multiple messages, separate them by new lines.
### Conversation Starters
Predefined clickable messages for user prompts. Separate multiple messages with new lines.
#### Chat Widget Pop Up
When enabled this will open the entire chat widget to the user on the webpage.
# AI Chat Widget
Customize the look and feel of each AI agent to match the branding of any business.
[Customize Chat ](/ai-features/ai-chat-widget/customize-chat-widget-appearance)
[Widget Apperance](/ai-features/ai-chat-widget/customize-chat-widget-appearance)[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/ai-agent-settings)
# Customize Chat Widget Appearance
This allows you to full customize the look, color, appearance and feel of the embeddable chat widget.
To customize the appearance of your AI agent's chat widget, click into any AI agent and then navigate to the Appearance tab. Here you will see a whole suite of options to customize pretty much every aspect of the chat widget.
On the right hand side there is a live preview of what your changes will look like.
### Font Size
Font size for the chat text in pixels.
### Font Family
Font family for the chat text.
### Subheading
Subheading for the chat widget, such as 'Online', 'Available', etc.
### AI Agent Icon
Displayed in the chat widget header as the AI Agent image.
Supported file extensions are png, jpg, jpeg and Ideal logo size is 32x32 pixels.
### Chat Input Placeholder
Enter the text that will be displayed as a hint in the chat input field.
### Send Message Button
Upload a custom send message button. Defaults to a standard button if not provided. Supported file extensions are png, jpg, jpeg and Ideal logo size is 32x32 pixels.
### Customize Colors
This section allows you to customize all the different colors of the platform.
### Widget Boarder
Border radius of the widget in pixels.
### Size of the Chat Widget
Adjust the size of the chat widget to small, medium or large.
### Tagline
### Chat Bubble Icon
This is the icon that appears on your site before a user clicks to open the chat.
### AI Agent Status Messages
Enable to show status messages from the AI agent in the chat widget which are displayed to the user while the AI is processing a response.
### Chat Avatar Icon
This image appears inside the chat window next to your bot’s messages.
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/ai-chat-widget)
# Chat Widget Voice / Audio
Information about how to enable voice in & out for the website chat widget.
#### What is Chat Widget Voice / Audio?
There is now the ability to have voice / audio functionality within your website chat widget.
This means that, when enabled there will be the option to click a voice icon and speak into the bot, as well as be able to hear the AI's response verbally.
Voice In = Voice to Text
Voice Out = Text to Voice
#### How to Enable Voice / Audio.
1. Navigate to the update tab for any AI agent
2. Click into the advanced settings
3. Enable/ Disable this feature based on what you want the user to be able to do
#### The Future of Voice / Audio
We will be implementing other native integrations that will allow for connections to OpenAI's Whisper and ElevenLabs. This will bring customized voices along with hundreds of higher quality options.
# Leads 2.0
Learn how to extract data from a conversation immediately, when a conversation ends or after 1-hour of no activity.
### Leads 2.0 Feature Documentation
Leads 2.0 is an improved system for extracting specific data from conversations with your AI agent. It allows you to collect structured information such as names, email addresses, appointment times, preferences, and other custom data points without requiring a separate prompt.
### When is Data Extracted?
Learn about the 3 different options for when data is extracted from the conversation.
#### 1. Immediate Data Extraction
* As soon as the user has answered all of the questions, the AI agent will immediately extract their responses and send the data to the notification email/ webhook.
* Based on the values you enter, the system will generate a prompt that you can copy and paste into your base system prompt.
* Replace the Custom\_Variables with the data you want to extract and add the following prompt to your base system prompt after AI agent data collection instructions:
Copy
```
## Data Collection Instructions:
After you have collected all of the data from the user: {Custom_Variable1, Custom_Variable2...}, call the send_notification_about_lead_collection() function.
```
#### 2. Conversation Ended
* As soon as a conversation has been ended by the user (refreshing the chat widget, starting a new conversation etc.) the system will review the entire conversation and extract the data values.
#### 3. After 1-Hour of No Activity
* If the AI agent is having a conversation with a user and the user doesn't respond for 1 hour, the AI agent will review the conversation and extract any available answers.
***
#### How to Use Leads 2.0
**Step 1: Configure Your Main System Prompt**
Include questions in your main system prompt that ask for the specific information you want to collect. These instructions should match the data values you enter in the leads tab to extract.
The base prompt tells the AI agent what questions to ask.
The Leads tab tells the AI agent which answers/data to extract.
Copy
```
(Example)
After answering the user's first question, ask them for their email address.
```
**Step 2: Define Data Extraction in the Leads Tab**
1. Navigate to the "Leads" tab in your AI agent settings
2. Define the data points you want to extract (e.g., email address)
3. Choose where to send the extracted data:
* Email notification
* Webhook URL (for CRM integration, Zapier workflows, etc.)
**Example of Data Points You Can Extract**
* Email addresses
* Names
* Phone numbers
* Appointment times
* Custom preferences
* Any specific information relevant to your business
**Additional Features**
* **Smart Extraction**: The system can intelligently extract information (e.g., extract a name from a personalized email address like "")
* **Partial Information Handling**: If a user only answers some of your questions, the system will still extract and forward the data they did provide
* **Data Storage**: All extracted lead data is stored within the system and visible in the conversations tab
* **CRM Integration**: Easily pipe extracted data to CRMs like Go High Level through webhooks
#### Compatibility Note
If you have existing AI agents, you'll see both "Legacy Leads" and "Leads 2.0" tabs. Existing agents will continue to function with the Legacy Leads system, while new agents can take advantage of the improved Leads 2.0 functionality.
#### Working with the Summary Feature
Leads 2.0 complements the existing Summary tab functionality, allowing you to extract both structured data (specific fields) and unstructured data (conversation summaries) from your AI interactions.
***
### Full Example Prompt
The example prompt below instructs the AI agent to answer the user's first question, then ask for their email address. After the user provides their email address, the AI agent will extract the data and send it to the saved webhook. After it sends that data the AI agent will ask the user if they want to schedule a meeting which then uses the scheduling prompt inside the scheduling tab.
Copy
```
// General Prompt Guidelines:
// When a line starts with two slashes like this line, it is considered a comment and will not be read by the AI agent.
// Replace the following content with your own company information or use one of the template prompts provided
// REMEMBER: Build your prompts iteratively (one change at a time). Make one change, and then test it. Etc...
// Prompts Should be written in English and then you can ask the AI agent to translate (If needed for your use case)
// This example handles all languages
// OR You can be more specific: This prompt is in English but I want you to interact with the users in Spanish
// Link based appointment scheduling (uncomment the below line, if needed)
// If a user wants to schedule a meeting or book an appointment, send them to this link: {YOUR_CALENDLY_LINK}
// Replace the placeholders in square brackets [] with your own information
## Basic Instructions:
- You are a helpful assistant for Agent Supply
- Your job is to answer questions customers send to you. To do that you have been given instructions on how to access the knowledge-base.
- If you do not have an answer to a question and it's not in the knowledge-base then let the user know that you do not have the answer to their question. You can say something like, "Hmm, I'm not sure."
- Keep your answers as concise as possible while still giving the required information.
- Do not break character. Avoid answering questions that are not at all relevant to the business.
- Speak to the User in the language they speak to you in.
// Data Collection Guidelines:
// Modify this section to match your use case and when you want the AI agent to ask the questions and collect the answers
// Make sure you tell the AI agent to ask the user for the all of info you have defined in the Leads data collection tab
// After the user has answered all of the questions, the AI agent will extract their answers and send the data via email or webhook
// Update the custom_variables to match the data values you want to extract from the conversation
// Note: If you are asking for the user's email, make sure you have defined 'email' in the leads tab as a piece of data to extract.
// The function prompt for immediate data extraction works best when the instruction is added right after the AI agent has asked the user the questions to collect the data.
## Lead Data Collection Instructions:
- Answer the first question from the user, then ask the user for their email address.
- After you have collected all of the data from the user: {email}, call the send_notification_about_lead_collection() function.
- After you have sent the lead collection data, offer the user the option to schedule a meeting.
```
***
#### Key Differences from Leads (Legacy - Leads 1.0)
| Feature | Legacy Leads | Leads 2.0 |
| ----------------------- | ----------------------------------------------- | --------------------------------------------------------------------------- |
| Prompt Requirement | Requires a custom prompt | Prompt is provided for you |
| Integration | Data extraction separate from main conversation | Data extraction integrated into main system prompt |
| Implementation | More complex setup | Simplified setup |
| Partial Data Collection | Limited handling of incomplete responses | Successfully extracts partial data even if users don't answer all questions |
# Leads (Legacy)
"Leads" allows you to extract data from users (like email address or phone number) when they interact with the AI Agent.
[Overview of Lead Generation](/ai-features/leads-legacy/overview-of-lead-generation)
[Lead Collection Form Fields](/ai-features/leads-legacy/lead-collection-form-fields)
[Lead Collection Webhook URL](/ai-features/leads-legacy/lead-collection-webhook-url)
[Lead Collection Email Notification](/ai-features/leads-legacy/lead-collection-email-notification)
# Overview of Lead Generation
Lead generation allows you to extract data from users when they interact with the AI Agent.
Within the context of Agent Supply there are 2 ways of generating leads.
#### 1. Collect in Form
Collect in Form before starting a conversation.
When you use this option a form will be shown to users that they will have to answer in order to start a conversation with a AI Agent.
#### 2. Collect in Conversation (Advanced)
Collect in chat during conversation with AI Agent
# Lead Collection Form Fields
Describing what each of the form fields represent.
### Predefined Fields
Create your own custom field OR pick a predefined field to auto-fill its corresponding values.
### Collection Method
Select the method used for collecting data.
#### In Form Collection
This will create a custom form that is shown to users before a conversation is started with a AI Agent.
#### In Chat Collection
This will train the AI Agent to ask the user these questions inside the chat conversation.
***
### Field Label
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/leads-legacy/overview-of-lead-generation)
# Lead Collection Webhook URL
When enabled, lead data will be sent to the specified URL upon collection.
After the AI Agent has collected all of the data from the user, it will compile all of that data together and then send it in a JSON format to the URL provided.
Example:
When this data is received by the CRM or other platform, the data will be formatted so you can easily extract custom values and match them with their custom fields in the client's CRM.
**1. Log into Agent Supply**
* Access your **Agent Supply** dashboard.
* Choose the specific **AI form or lead generation tool** you're using to collect user data.
**2. Open Integration Settings**
* Within Agent Supply, navigate to the **Integrations** or **Webhook/API** settings section for the form.
* This is where you configure where the lead data should go.
**3. Copy the Webhook URL from GoHighLevel**
* Open your **GoHighLevel** dashboard.
* Go to **Settings → Webhooks** (or **Triggers**, depending on setup).
* Create or retrieve a **Webhook URL** that will receive incoming lead data.
* Copy that URL.
**4. Paste Webhook into Agent Supply**
* Back in Agent Supply, paste the **GHL Webhook URL** into the webhook destination field.
* This tells Agent Supply to send new lead submissions to GoHighLevel.
**5. Map Form Fields**
* Match each form field from Agent Supply to its corresponding field in GoHighLevel:
* First Name → First Name
* Email → Email
* Phone → Phone
* And so on...
* This ensures the data is correctly placed in the GHL contact record.
**6. Submit a Test Lead**
* Use your Agent Supply form to submit a **test lead**.
* Then check your **GoHighLevel CRM** to verify the contact appears correctly with the expected data.
**7. Troubleshooting**
If the test doesn't work:
* Double-check the **Webhook URL** is correct and live.
* Make sure **field names are mapped properly**.
* Look at any **logs or error messages** in Agent Supply or GHL to diagnose issues.
***
#### 🔁 Why Use This Integration?
* **Saves Time**: No more copying and pasting lead info.
* **Improves Accuracy**: Automation avoids typos and delays.
* **Enables Automation**: As soon as the lead hits GoHighLevel, you can trigger emails, SMS, calls, or workflows.
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/leads-legacy/lead-collection-form-fields)
# Lead Collection Email Notification
When enabled, lead data will be sent to the provided email upon collection.
When a new lead is collected, we will send an email to this email address.
[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/leads-legacy/lead-collection-webhook-url)
# Sending Lead Data to Zapier
### 🔁 Objective:
Send **Lead data** (like name, email, phone, tags, etc.) from **Agent Supply** to **Zapier**, enabling you to automate actions such as CRM updates, email follow-ups, or Slack notifications.
***
### ✅ Prerequisites:
1. A **Zapier** account
2. Access to your **Agent Supply dashboard**
3. An **Agent Supply assistant** that collects lead data (e.g., via form, user message, or custom flow)
***
### 🧩 Step-by-Step Integration (Zapier Webhook)
#### **Step 1: Create a Zap in Zapier**
1. Log in to [Zapier.com](https://zapier.com)
2. Click **“Create Zap”**
3. For the **Trigger**, choose:\
→ **Webhooks by Zapier**\
→ Event: **Catch Hook**\
→ Click **Continue**
#### **Step 2: Copy Your Webhook URL**
1. Zapier will provide a **custom webhook URL**\
(e.g., `https://hooks.zapier.com/hooks/catch/123456/abcde`)
2. **Copy** this URL for use in Agent Supply
***
### 🔗 Step 3: Set up a Webhook in Agent Supply
#### Option A: Use an **Action Step** inside your Agent Supply Assistant
If you're using an agent with defined actions or a flow:
1. Go to your agent’s flow in the Agent Supply dashboard.
2. Locate the step where the lead is captured.
3. Add an **"HTTP Request"** or **Webhook** step.
4. Set the method to **POST**
5. Paste in the **Zapier Webhook URL**
6. Structure the JSON payload. Example:
```json
jsonCopiarEditar{
"name": "{{user.name}}",
"email": "{{user.email}}",
"phone": "{{user.phone}}",
"tags": "{{user.tags}}"
}
```
*(Replace `{{user.variable}}` with the actual variable names used in your assistant.)*
7. Save and deploy the flow.
#### Option B: Use Global Webhooks (if Agent Supply provides this)
Some Agent Supply projects allow **global webhook configuration**:
1. Navigate to project settings or integrations.
2. Look for “Webhook” or “Lead Forwarding” settings.
3. Paste the Zapier Webhook URL.
4. Define what event will trigger the webhook (e.g., “on lead creation”).
5. Customize the payload if needed.
***
### 🧪 Step 4: Test the Integration
1. Return to your Zap in Zapier.
2. Click **“Test Trigger”**
3. Submit a lead via your Agent Supply assistant (preview or live).
4. Zapier should receive the lead data.
5. Click **Continue** when the test succeeds.
***
### ⚙️ Step 5: Add Action Steps in Zapier
Once Zapier has received the lead data, you can choose what happens next:
* Send it to a CRM (e.g., Salesforce, Pipedrive, HubSpot)
* Log it in a spreadsheet
* Add to your email list
* Send internal alerts (e.g., Slack or Gmail)
Example:
1. Choose **“Create Contact in HubSpot”**
2. Map the lead fields (e.g., email → email)
3. Save and **turn your Zap on**
***
### ✅ Integration Complete!
Now each time a lead interacts with your **Agent Supply assistant**, the captured data is instantly sent to Zapier, triggering your custom workflow.
# Sending Lead Data to GHL
### ✅ Goal:
Send lead data (e.g., name, email, phone) from **Agent Supply** (via an Agent or Form) to **GoHighLevel (GHL)**, so it appears as a Contact or triggers a Workflow in GHL.
***
### 🔧 Step-by-Step: Send Lead Data from **Agent Supply** to **GoHighLevel**
***
#### **Step 1: Prepare Your GoHighLevel Webhook URL**
1. **Log in to your GoHighLevel account**.
2. Navigate to:
* **Settings** > **Company Settings** > **API Keys** → *Copy your API key*.
3. To trigger a workflow via webhook:
* Go to **Workflows**.
* Click **+ Create Workflow** → choose **Start from Scratch**.
* Add a **Webhook Trigger**.
* Copy the **Webhook URL** generated — you'll need this in Agent Supply.
***
#### **Step 2: Go to Your Agent in Agent Supply**
1. In **Agent Supply**, open the **Agent** (bot) where you’re capturing leads.
2. Insert a **"Send Webhook"** action after the data is collected (e.g., after a form or dialogue).
3. Paste the **GoHighLevel Webhook URL** into the webhook step.
***
#### **Step 3: Format the Payload Correctly for GHL**
Use this JSON structure in the webhook action:
```json
jsonCopiarEditar{
"firstName": "{{lead.firstName}}",
"lastName": "{{lead.lastName}}",
"email": "{{lead.email}}",
"phone": "{{lead.phone}}",
"source": "AgentSupplyBot",
"customField": "value"
}
```
> Replace variables with the fields your **Agent Supply agent** collects.
***
#### **Step 4: Test the Webhook from Agent Supply**
1. In Agent Supply, use the **“Test Webhook”** feature with sample data.
2. Go to GoHighLevel:
* Check **Contacts** to confirm receipt of the lead.
* If you’re triggering a workflow, go to **Workflow → History** to see if the trigger fired.
***
#### **Step 5: Optional – Use the GHL API for Direct Integration**
If you want to **skip the GHL webhook trigger** and directly create a contact, use an **API call** in Agent Supply:
**Endpoint**:
```
bashCopiarEditarhttps://rest.gohighlevel.com/v1/contacts/
```
**Headers**:
```json
jsonCopiarEditar{
"Authorization": "Bearer YOUR_API_KEY",
"Content-Type": "application/json"
}
```
**Body**:
```json
jsonCopiarEditar{
"firstName": "John",
"lastName": "Doe",
"email": "john@example.com",
"phone": "+1234567890",
"locationId": "YOUR_LOCATION_ID",
"customField": "value"
}
```
In Agent Supply:
* Use the **API Call** block.
* Set method to **POST**, enter the URL, headers, and body.
* Make sure your data variables (e.g., `{{lead.email}}`) are correctly mapped.
***
### 🧠 Pro Tips
* Use **Test Mode** in GoHighLevel to verify the trigger and payload.
* Validate emails and phone numbers in Agent Supply before sending.
* Use **custom fields** in GHL to capture additional info like UTM tags, lead source, etc.
* You can chain multiple actions in Agent Supply: e.g., save to Airtable + send to GHL.
# Sending Lead Data to Zoho CRM
Tutorial on how to setup a webhook from Zoho and connect to your Agent Supply account.
#### ✅ Prerequisites
Make sure you have:
1. A **Zoho CRM account** with API access enabled.
2. A valid **OAuth 2.0 access token** from Zoho (guide below).
3. An **Agent Supply account** with access to build **Chains** (Agent workflows).
***
#### 🧩 Step-by-Step: Send Lead Data to Zoho CRM via Agent Supply
***
**Step 1: Generate Your Zoho CRM API Access Token**
Zoho uses OAuth 2.0. Here’s how to get the token:
1. Go to Zoho API Console.
2. Create a **new client** (select **Server-based Applications**).
3. Set a **redirect URI** (it can be dummy if using the "Self Client").
4. Generate a code using the **"Self Client"** tab.
5. Exchange the code for an access token using this command:
```bash
bashCopiarEditarcurl --request POST \
--url https://accounts.zoho.com/oauth/v2/token \
--data 'grant_type=authorization_code&client_id=YOUR_CLIENT_ID&client_secret=YOUR_CLIENT_SECRET&redirect_uri=YOUR_REDIRECT_URI&code=YOUR_CODE'
```
Save the `access_token` and `refresh_token` securely.
***
**Step 2: Capture Lead Data in Agent Supply**
1. Go to the **Chain Builder** in Agent Supply.
2. Add steps to collect user data:
* Name
* Email
* Company
* Phone
3. Store each piece of data in variables (e.g., `{{first_name}}`, `{{email}}`, etc.).
***
**Step 3: Add HTTP Request Node to Your Chain**
1. Insert an **HTTP request** node after capturing lead data.
2. Configure it with the following:
**🌍 URL**
```
rubyCopiarEditarhttps://www.zohoapis.com/crm/v2/Leads
```
**🔧 Method**
```
nginxCopiarEditarPOST
```
**🔐 Headers**
```json
jsonCopiarEditar{
"Authorization": "Zoho-oauthtoken YOUR_ACCESS_TOKEN",
"Content-Type": "application/json"
}
```
**🧾 Body (JSON)**
Fill in Zoho CRM fields using your stored variables:
```json
jsonCopiarEditar{
"data": [
{
"Company": "{{company}}",
"Last_Name": "{{last_name}}",
"First_Name": "{{first_name}}",
"Email": "{{email}}",
"Phone": "{{phone}}",
"Lead_Source": "Agent Supply"
}
]
}
```
> Replace variable names with whatever you're using in your Chain.
***
**Step 4: Test Your Chain**
* Run a test session through your Chain.
* Submit lead data.
* Go to **Zoho CRM > Leads** to confirm the lead was created.
***
**Step 5: Handle Token Expiry (Optional)**
Access tokens expire (usually after 1 hour). To handle this:
* Use your `refresh_token` to generate new tokens on demand.
* You can set up a separate Agent Supply Chain to refresh the token:
```bash
bashCopiarEditarcurl --request POST \
--url https://accounts.zoho.com/oauth/v2/token \
--data 'refresh_token=YOUR_REFRESH_TOKEN&client_id=YOUR_CLIENT_ID&client_secret=YOUR_CLIENT_SECRET&grant_type=refresh_token'
```
Store and reuse the new access token in your Agent Supply Chain.
***
#### ✅ Summary
| Step | Description |
| ---- | ---------------------------------------------- |
| 1 | Get Zoho OAuth access token |
| 2 | Collect lead data in Agent Supply Chain |
| 3 | Use HTTP request node to send data to Zoho CRM |
| 4 | Test and confirm leads appear in Zoho |
| 5 | (Optional) Automate token refresh |
####
# Scheduling
There are a few options for how you can your AI agent to schedule meetings.
The easiest and most reliable way to get more meetings scheduled is to train the AI agent to provide a calendar link that users can click and schedule a meeting.
[AI Scheduling - Simplified](/ai-features/scheduling/ai-scheduling-simplified)
The more advanced option is to enable in-chat scheduling. This will connect your AI agent directly to a Google calendar and work with the users to find the best available time. Please note, that this is an experimental custom function. Changing or updating the prompt may cause scheduling to not work.
[AI Scheduling - Experimental (Beta)](/support/resources/sales-deck/ai-agency-101/ai-saas-funnel-explained)
# AI Scheduling - Simplified
The easiest way to create an AI agent that can schedule appointments is to provide a calendar link when appropriate.
### Base System Prompt for Scheduling
In your base system prompt simply add in a line that states:
```
If a user wants to schedule a meeting or book an appointment,
send them to this link: https://yourlink
```
When a user wants to schedule a meeting they will be able to click the link and book the meeting.
For more advanced in-chat scheduling functionality go here:
[AI Scheduling - Experimental (Beta)](/ai-features/scheduling/ai-scheduling-experimental-beta)[ ](https://docs.stammer.ai/stammer.ai-docs/ai-features/scheduling)
# AI Scheduling - Experimental (Beta)
This is an advanced, custom function for AI agents to execute and is in beta. Use with caution.
This allows your A.I. agent to see your current availability via your connected calendar and then be able to automatically schedule appointments directly to the calendar.
[What is Scheduling?](/ai-features/scheduling/ai-scheduling-experimental-beta/what-is-scheduling)[Calendar Integration](/ai-features/scheduling/ai-scheduling-experimental-beta/calendar-integration)[Appointment Scheduling Activation](/ai-features/scheduling/ai-scheduling-experimental-beta/appointment-scheduling-activation)[Appointment Scheduling Webhook URL](/ai-features/scheduling/ai-scheduling-experimental-beta/appointment-scheduling-webhook-url)[Appointment Scheduling Email Notification](/ai-features/scheduling/ai-scheduling-experimental-beta/appointment-scheduling-email-notification)[Appointment Scheduling Base System Prompt](/ai-features/scheduling/ai-scheduling-experimental-beta/appointment-scheduling-base-system-prompt)
# What is Scheduling?
Our AI Agent now offers a convenient appointment scheduling feature. This feature integrates seamlessly with Google Calendar to give you real-time availability slots. Here's how you can make the most out of it:
**Setting Up:**
1. Turn on the scheduling feature using the provided switch/button.
2. Connect your Google Calendar. The system will guide you through a secure OAuth process to ensure your data remains private.
3. Set your working hours. In the initial version, set a uniform time for all days. For example, you might work 9 a.m. to 5 p.m. This will be displayed in your time zone, but stored in UTC for consistency.
**Using the Feature:**
1. When a user wishes to schedule an appointment, the AI Agent displays available times based on your Google Calendar.
2. Once a time is chosen, the AI Agent verifies its availability.
3. The user provides their email address, and the AI Agent schedules the appointment. It can also set up a Google Hangout Meet automatically.
4. Both you and the user receive a notification through Google's default system once the appointment is booked.
**Additional Info:**
Remember, the AI Agent requires the user's email to schedule an appointment. This ensures you can send a Google Meet invite if needed.
# Calendar Integration
# Google Calendar
The Google calendar integration allows you to connect your Google calendar directly to an A.I. Agent.
We created a standalone brandless integration called "AI Meeting Scheduler". This allows you or your clients to connect their Google calendar to their AI Agent without seing anything related to Agent Supply.
By default the AI Agent will only look at the main calendar. If you have other calendars inside your Google account, make sure the main calendar is able to see all of the availability.
# Appointment Scheduling Activation
This is how you enable in-chat scheduling for your AI agent.
**How Appointment Scheduling Works**
**Overview:**
Our AI agent now offers a convenient appointment scheduling feature. This feature integrates seamlessly with Google Calendar to give you real-time availability slots. Here's how you can make the most out of it:
**Setting Up:**
1. Turn on the scheduling feature using the provided switch/button.
2. Connect your Google Calendar. The system will guide you through a secure OAuth process to ensure your data remains private.
3. Set your working hours. In the initial version, set a uniform time for all days. For example, you might work 9 a.m. to 5 p.m. This will be displayed in your time zone, but stored in UTC for consistency.
**Using the Feature:**
1. When a user wishes to schedule an appointment, the AI agent displays available times based on your Google Calendar.
2. Once a time is chosen, the AI agent verifies its availability.
3. The user provides their email address, and the AI agent schedules the appointment. It can also set up a Google Hangout Meet automatically.
4. Both you and the user receive a notification through Google's default system once the appointment is booked.
**Additional Info:**
Remember, the AI agent requires the user's email to schedule an appointment. This ensures you can send a Google Hangout Meet invite if needed.
# Appointment Scheduling Webhook URL
Stay up to date and notified about all new appointments.
Enter in a webhook URL where you want data to be sent.
For example, enter in a webhook URL from Zapier or your CRM and every time a new meeting is scheduled the data below will be sent.
If you don't want to use the webhook, you can use the email notification.
# Appointment Scheduling Email Notification
Send an email every time there is a new appointment scheduled.
When a new appointment is scheduled, we will send an email to this email address.
# Appointment Scheduling Base System Prompt
The appointment scheduling base system prompt is the special prompt that is appended to your main base system prompt to auto-train your AI agent to know how to schedule an appointment.
If you know what you are doing, you can modify this additional prompt as needed.
# The Appointment Scheduling Functions & Customizing the Appointment Scheduling Prompt
This page explains the appointment scheduling functions that we provide to the AI agents and also gives some tips on customizing the appointment scheduling prompt.
### Intro
Appointment scheduling in Agent Supply is entirely driven by the prompts you write for your AI Agent This tutorial will walk you through the process of customizing the appointment scheduling prompt to fit your specific requirements.
### Example Appointment Scheduling Prompt
```
// ...Rest of your prompt above...
You are an appointment scheduling bot that answers questions. You will always prompt the user to see if they want to schedule an appointment.
// Change the above if your scenario is different
# How to schedule an appointment
## Get the timezone from the user
When the user wants to schedule an appointment you will ask them what timezone they are in.
## Get the available times
Then you will use the get_available_times function, passing the user's timezone converted by you to a valid IANA Time Zone string, to get the available times for the agent.
Example call: get_available_times(timezone="America/Los_Angeles")
## Propose the times to the user
Then I want you to propose the next 3 time slots the agent is available in the following format:
1. Mon 11-04-2024 11:00AM PST
2. Mon 11-04-2024 11:30AM PST
3. Mon 11-04-2024 2:00PM PST
Show these timeframes using markdown for clarity. Each day and its available time should be displayed on separate lines for easy readability.
If the user doesn't like any of those times give them 3 others that will work.
## Scheduling the appointment
When the user picks a time, you will then get their email. If user has already provided their email, then ask if they want to use the same one, else ask for their email.
Once you have their email and the time you will then use the schedule_appointment function to schedule the appointment by calling it with the email and the time in RFC 3339 format.
Example call: schedule_appointment(email="someone@example.com", desired_time_rfc3339="2024-11-04T09:30:00-08:00")
Important Note: This is just an example call for you to understand, please do not use this email, instead, ask the user for their email and use that.
## Informing the user the appointment has been scheduled
When the appointment has been successfully scheduled, go ahead and include the link that the function gives you to the successfully scheduled appointment.
```
However, occasionally users do want to modify or customize their appointment scheduling prompt.
### Note on the "Appointment Scheduling Base System Prompt"
In Agent Supply, there is an "Appointment Scheduling Base System Prompt" that gets appended to the main base system prompt. This separation allows for clearer organization, but you can choose to merge the appointment scheduling prompt into your main base system prompt if desired.
### Customizing the Appointment Scheduling Prompt
If you do want to customize the appointment scheduling prompt, you will need to know about the two appointment scheduling functions we provide to the AI so that it knows how to call our code and get the function scheduled.
#### The get\_available\_times function
```
get_available_times(timezone="America/Los_Angeles")
```
The get\_available\_times function takes in a [valid IANA timezone string](https://en.wikipedia.org/wiki/List_of_tz_database_time_zones) and returns the a list of time ranges that the user whose calendar is connected to the agent is available.
**Example Output of the function**
```
Thu 04-04-2024 06:00PM-05:00PM PDT
Fri 04-05-2024 12:00PM-01:00PM PDT
Fri 04-05-2024 03:00PM-05:00PM PDT
Mon 04-08-2024 12:00PM-01:00PM PDT
Mon 04-08-2024 03:00PM-05:00PM PDT
Tue 04-09-2024 12:00PM-01:00PM PDT
Tue 04-09-2024 03:00PM-05:00PM PDT
Wed 04-10-2024 12:00PM-01:00PM PDT
Wed 04-10-2024 03:00PM-05:00PM PDT
```
#### The schedule\_appointment function
```
schedule_appointment(email="someone@example.com", desired_time_rfc3339="2023-11-04T09:30:00-08:00")
```
This is the actual function that schedules the appointment on the calendar of the user who is connected to the AI Agent.
It takes in two parameters:
* The email of the person chatting with the agent who wants to schedule the appointment
* The desired time to schedule the appointment in RFC 3339 format.
**Example Output of the function**
```
Appointment scheduled successfully. Event Link: https://www.google.com/calendar/event?eid=YmwxNjM1cjhjaGRqdWgkeajlkgjreNGtpZzAgdGl
```
# Integrations
This is how you can integrate AI Agents into any business. The most common and popular integration that 99% of agencies and businesses use is the option to embed it on a website.
[Embed AI Agent on a Website](/ai-features/integrations/embed-ai-agent-on-a-website)
There are also native integrations for certain social media channels which allow you to place an A.I. Agent on anywhere.
[Instagram DM](/ai-features/integrations/instagram-dm)[Facebook Messenger](/ai-features/integrations/facebook-messenger)[WhatsApp](/ai-features/integrations/whatsapp)[Zapier App](/ai-features/integrations/zapier-app)
# Embed AI Agent on a Website
How to embed an AI agent on a website
To embed an AI Agent on a website, you can go to any AI Agent in your dashboard and go to the Integrations tab to see the embed code and available settings to control when the AI Agent pops up.
## Squarespace
### 🔧 How to Embed the Agent Supply Chat Widget into a Squarespace Site (Step by Step)
***
#### 1. **Introduction**
This guide walks you through embedding the **Agent Supply** chat widget into your Squarespace site so visitors can chat with your AI or support agent directly.
***
#### 2. **Obtain the Widget Code**
* Log into your **Agent Supply** account (or create one).
* Go to your dashboard and find the **embed code** for your chat widget.
* This is usually a JavaScript or HTML snippet meant to be copied and embedded into your website.
***
#### 3. **Access Squarespace Editor**
* Open your Squarespace website and navigate to the page where you'd like the chat to appear.
* Click **Edit** to enter the page builder.
***
#### 4. **Insert a Code Block**
* Click the **+** button to add a new block to the page.
* Select the **“Code”** block.
* Paste the **Agent Supply** chat widget code into the code block.
* You can place this in the page footer or any section where you want the widget to appear.
***
#### 5. **Save and Publish**
* Click **Apply** or **Save** after pasting the code.
* Exit the editor and **publish** the page to make your changes live.
***
#### 6. **Verify the Widget on the Live Site**
* Visit your live Squarespace site in a regular browser window.
* You should see the **Agent Supply** chat widget — usually a floating icon in the bottom corner.
* Open it to make sure it works and is connected to your **Agent Supply** account.
***
#### 7. **Troubleshooting (if the widget doesn’t show up)**
* Make sure:
* You used a **Code** block (not text).
* The script was pasted correctly.
* You saved and published changes.
* There are no conflicting scripts.
* You’ve cleared your browser cache.
***
#### 8. **Conclusion**
Your site now has the **Agent Supply** chat widget active, enabling real-time AI-powered chat. You can further customize its appearance or behavior in your **Agent Supply** dashboard.
***
#### ✅ Summary Table
| Step | Action |
| ---- | -------------------------------------------- |
| 1 | Copy chat embed code from **Agent Supply** |
| 2 | Open Squarespace editor on your desired page |
| 3 | Add a **Code** block and paste the code |
| 4 | Save and publish |
| 5 | Verify the chat is working on the live site |
## Shopify
### ✅ **How to Embed the Agent Supply Chat Widget into a Shopify Store**
#### 1. **Introduction & Overview**
* The goal is to enable real-time AI Agent interaction on your storefront using **Agent Supply**.
***
#### 2. **Accessing Your Shopify Theme Code**
* In your Shopify admin panel, go to **Online Store → Themes**.
* Find your current active theme and click **“Actions” → “Edit code”**.
***
#### 3. **Opening the Correct File**
* In the left sidebar, open the **`theme.liquid`** file under the **Layout** section.
* This file contains your site’s overall HTML structure — the place to insert the widget code.
***
#### 4. **Inserting the Agent Supply Widget Code**
* Copy the **Agent Supply chat widget script** (usually a `
