How to Build and Manage AI Agents in myCRMSIM

The myCRMSIM AI Agent Builder allows you to create, test, manage, and automate AI Agents for customer conversations, lead qualification, appointment booking, support handling, and CRM workflows.
AI  Agents can be used for different business purposes such as sales, customer support, appointment setting, product enquiries, franchise enquiries, lead follow-up, and automated conversation handling.
This guide explains how to use the AI Agents Dashboard, test an AI Agent, create a new AI Agent, configure each setup step, connect automations, and activate the agent properly.

To access AI Agents, go to the left sidebar and click:
AI Dashboard
The AI Agents Dashboard is the main page where you can view, create, test, activate, pause, edit, duplicate, export, and delete AI Agents.
This dashboard helps you manage your AI workforce from one place.

Before creating an AI Agent, make sure the OpenAI API Key is already configured in the system.
When you click on Create New AI Agent, the system requires a valid API key to continue the AI agent setup. If the API key is not configured, the AI agent may not be created or may not work properly.
Please ensure the OpenAI API Key is set before starting the AI Agent creation process.

On the AI Agents Dashboard, all created AI Agents are displayed in a table.
Each row represents one AI Agent.
The dashboard includes the following main columns:

The Agent Name shows the name of the AI Agent or strategy.
This name helps you quickly identify which agent is being used for which purpose.

The Role column shows the job or responsibility assigned to the AI Agent.
The role helps define how the AI Agent should behave during conversations.

The Persona column shows a short preview of the agent’s instructions or personality.
This gives you a quick idea of how the AI Agent has been trained to speak, respond, and guide customers.

The GHL Tag column shows the tag connected with the AI Agent.
Tags are important for CRM tracking because they help organize contacts and trigger the correct automation workflows.

The Model column shows which AI model is being used by the agent.
You can also change the model from the dropdown.
Available model options may include:
GPT-4o mini
GPT-4.1 mini
GPT-4.1
GPT-5 mini
GPT-5
GPT-5.1
Choose the model based on your required speed, cost, and response quality.

The Status column allows you to turn an AI Agent ON or OFF.
When the toggle is ON, the agent is active.
When the toggle is OFF, the agent is paused.
This is useful when you want to temporarily stop an agent without deleting it.

The Actions column gives quick management options for each AI Agent.
Available actions include:

Downloads the agent strategy as a JSON file for backup or reuse.

Opens the AI Agent Builder so you can update the agent setup.

Removes the AI Agent from the dashboard.
Before deleting an important agent, it is recommended to export a JSON backup.

The filter dropdown allows you to filter the list of agents.
Example:
All
Use this option when you want to view all agents or narrow down the list.

The search bar allows you to search for an agent by name or keyword.
This is useful when your workspace has multiple AI Agents and you want to find a specific one quickly.

The Test AI button allows you to test an AI Agent before using it with real customers.
This is a safe testing area where you can simulate a conversation with your selected AI Agent. The test session is isolated, which means it will not affect live leads, real contacts, or active customer conversations.
To open the simulator, click the Test AI button from the AI Agents Dashboard.

After clicking Test AI, a side panel opens on the right side of the screen.
This panel is called:
Test AI Simulation
At the top, you will see a session ID. This identifies the current test session.
You may also see a message such as:
Please select an agent to begin simulation.
This means you need to select an AI Agent before starting the test conversation.

The left side of the Test AI panel includes the Simulator Settings section.
This section allows you to choose which AI Agent and contact should be used for the test.

The Agent dropdown is used to select the AI Agent you want to test.
After selecting the agent, the simulator uses that agent’s instructions, tone, model, conversation flow, configurations, and knowledge.
Examples:
Plant Seller

The Contact dropdown is used to select a contact for the simulation.
This helps test how the AI Agent behaves when speaking with a selected contact.
It is useful for testing contact-based behavior, personalization, and CRM-related flows.

The Timezone field shows which timezone is being used for the test session.
Example:
Asia/Karachi
This is useful when testing execution windows, response timing, and business-hour behavior.

Click New Conversation to start a fresh test conversation.
Use this option when you want to clear the previous test and begin again from the start.
This is helpful when testing multiple scenarios such as interested leads, not interested leads, booking requests, objections, or stop messages.

The Chat History section shows previous test conversations for the selected agent.
If no agent is selected, you may see:
Select an agent to see history
Once an agent is selected and tested, chat history can help you review previous test conversations.

The right side of the Test AI panel is the main chat testing area.
Before selecting an agent, you may see:
Select an agent and start chatting
After selecting the agent and contact, you can type a test customer message and start the conversation.
The message box is available at the bottom of the simulator. Type your message and click Send or press Enter.
Example test message:
Hi, I am interested in your service.
The AI Agent will respond based on its setup.

To create a new AI Agent, visit the Ai Dashboard from side menu and click the blue button named as:
+ Create New AI Agent
You can see it at the top right of the AI Agents Dashboard.
After clicking this button, a popup will appear titled:
Choose Your Path
This popup allows you to select how you want to create your AI Agent.

Select Create New if you want to build a new AI Agent from scratch.
Use this option when you want to manually create the agent’s:
1. Instructions
2. Conversation flow
3. Configurations
4. Knowledge
5. Tasks
This is the best option when setting up a completely new AI Agent.

Select Import Strategy if you already have an exported AI Agent strategy file.
This option allows you to upload a .json file and recreate an existing strategy.
Use this option for importing a backup, reusing a previous strategy, moving an agent setup from another workspace, or restoring a saved configuration.

For users who prefer a guided setup, myCRMSIM offers “Simi,” an AI assistant designed to help
you build your agent conversationaly.

1. Click + Create New AI Agent.
2. Select Create with Simi and click Next.
3. Click the Start with Simi button.
4. A chat interface will open where you can describe what you want the agent to do. 

For example, you can ask, “how can i create an agent?” and Simi will guide you through the required parameters

 

After selecting the creation method, the AI Agent Builder will guide you through the setup process.
The builder includes five main steps:
Step 1: Instructions
Step 2: Conversation
Step 3: Configurations
Step 4: Knowledge
Step 5: Tasks
Each step controls a different part of how the AI Agent works.

The Instructions step is where you define who the AI Agent is, what role it should perform, how it should speak, and which AI model it should use.
This is one of the most important steps because it defines the identity and behavior of the agent.

The Strategy Name is the internal name of your AI Agent.
Use a clear name that explains the purpose of the agent.

The Agent Role defines what the AI Agent should act as during the conversation.

The Contact Tag is used to apply a specific tag to contacts connected with this AI Agent.
Tags are important because they help organize contacts inside your CRM.
Example tags:
qualified_sales_lead
plant_lead
Using the correct tag helps you filter contacts, trigger workflows, and track lead status properly.

The Tone controls how the AI Agent communicates with customers.
Choose a tone that matches your business and the type of conversation.

Best for corporate, B2B, legal, finance, or serious business conversations.

Best for general customer support and simple everyday conversations.

Best for sales, promotions, offers, and lead engagement.

Best for customer support, complaints, service issues, or sensitive conversations.

Best for product education, onboarding, guides, and FAQs.

Best for sales follow-ups, appointment booking, and lead conversion.

Best for support conversations where the customer needs confidence and clarity.

The AI Model section allows you to select which AI model will power the agent.
Different models may vary in quality, speed, reasoning ability, and cost.
Available model options may include:

Fast and cost-effective. Suitable for most simple sales, support, and lead qualification conversations.

A lightweight model option for faster and lower-cost conversations.

A stronger model for better reasoning and more detailed responses.

A faster GPT-5 model option for higher-volume conversations.

A powerful model for advanced reasoning and high-quality conversations.

Best for more complex AI Agent workflows, premium conversations, and advanced reasoning tasks.

Choose the model based on the quality, speed, and cost requirements of your agent.

Once you have completed all the sections, click the Save Compilation button at the bottom right of the modal.
The Prompt Builder will close, and you will be returned to the Strategy Builder page. Your newly created, fully structured prompt will now be displayed in the Prompt Editor panel, ready to be used by your AI agent! You can always click "Edit Instructions" if you need to make changes later.

The Advanced Settings section is optional and allows you to control how the AI agent responds during conversations.
Click on Advanced Settings to expand this section. Inside this section, you can configure the following options:

The Initial Message Delay controls how long the AI waits before sending the first message.
Options may include:
Instant
30 seconds – 1 minute
1 – 5 minutes
5 – 15 minutes
15 – 30 minutes
You can also set custom minimum and maximum delay values.
This makes the AI feel more natural instead of responding instantly every time.

The Response Delay controls how long the AI waits before replying after the customer sends a message.
Options may include:
Instant
5 – 15 seconds
15 – 30 seconds
30 – 60 seconds
1 – 2 minutes
A small response delay can make the conversation feel more human and realistic.

The Objection Handling setting controls how the AI responds when a customer hesitates, says no, asks difficult questions, or gives an objection.
The behavior may be adjusted between:
Passive
Persuasive
Confident
Assertive
For sales agents, a persuasive or confident style is usually helpful.
For support agents, a calm and supportive style is usually better.

The Qualification Priority setting controls how strictly the AI qualifies a lead before moving the conversation forward.
The behavior may be adjusted between:
Light
Balanced
Strict
Use Light when you want the AI to keep the conversation going and avoid being too strict.
Use Balanced for normal lead qualification.
Use Strict when only serious and qualified leads should continue.

The Creativity setting controls how freely the AI responds during the conversation.
The behavior may be adjusted between:
Scripted
Adaptive
Freeform
Use Scripted when you want the AI to stay close to the prompt.
Use Adaptive when you want natural but controlled replies.
Use Freeform when you want the AI to respond more flexibly and creatively.

The Conversation step is used to build the AI Agent’s complete message flow. In this step, you can set the first message, add qualification questions, create follow-ups, use templates, and add conditional logic.
This step helps the AI guide the customer in a structured way and qualify the lead properly.

The center area shows the full conversation flow visually.
The flow starts from:
Conversation Start
Then moves to:
Initial Message
After that, the flow continues through questions, follow-ups, and conditions.
You can use the canvas to review how the AI conversation will move from one step to another.

The Initial Message is the first message the AI Agent sends when the conversation starts.
This message should be short, clear, and relevant to the customer.

Hi, thanks for reaching out. How can I help you today?
A good initial message should quickly explain what the AI can help with and encourage the customer to reply.

The Spintext option allows you to add multiple variations of the initial message.
You can add up to 5 variations, and the AI will rotate between them for new leads.
This helps the AI sound more natural and prevents every customer from receiving the exact same message.

Hi, thanks for reaching out. How can I help you today?

Hello, I’m here to help. What are you looking for?

Hi, welcome. Are you interested in learning more or booking an appointment?

Using Spintext makes the conversation feel less robotic and more personalized.

The Qualification toggle controls whether the AI should qualify the lead during the conversation.
When qualification is enabled, you can add questions that help the AI understand the customer’s interest, need, budget, or intent.
This is useful for sales, appointment booking, and lead filtering.

Use Add Question to add qualification questions into the conversation flow.
Each question appears as a separate node in the flow.

Are you mainly looking to get new leads, keep current clients engaged, or both?
Are you currently running any marketing campaigns?
What does your ideal customer look like?
On the right-side panel, you can edit the question text, move the question up or down, delete the question, or enable/disable qualification for that question.

Conditional Logic allows you to create different conversation paths based on the customer’s answer.

If the customer answers yes, the AI can move to the next question or skip to another step.
If the customer answers no, the AI can send a different message or continue with another question.
Use Add Condition to define these rules.
Conditional logic helps make the conversation smarter and more relevant to the customer’s reply.

Use Add Follow-up to add a follow-up message after a question or conversation step.
Follow-ups are useful when the customer does not reply or when you want to remind them after some time.
You can set the follow-up message and delay time using:
Days
Hours
Minutes

Send a follow-up after 1 day if the customer does not respond.

The Follow-up Spintext option allows you to add variations of a follow-up message.
This helps the AI avoid sending the same follow-up every time.
It makes follow-up messages feel more natural and less repetitive./

Use Message Template to quickly insert a pre-written message.
Templates help save time and keep your messaging consistent.

Confirmation message
Intro message
Value teaser message
Booking message
Select a template and add it to the conversation flow.

Use Follow-up Template to quickly add a pre-written follow-up message.
This is useful when you want to add standard follow-ups such as:
Checking in
Reminder message
Offer reminder
Booking reminder
Templates make it easier to build the conversation flow faster.

Each message, question, or follow-up node can be edited or deleted.
Use the edit option to update the text.
Use the delete option if the step is no longer needed.
This allows you to improve the conversation flow at any time.

The Configurations step allows you to control how the AI Agent sends messages, what goal it should work toward, how it should handle unqualified leads, and when it is allowed to message contacts.
This step is important because it helps the AI Agent follow your business rules.

The Messaging section controls how the AI formats its replies.

The Multiple Messages option allows the AI to split long responses into shorter sequential messages.
When this option is turned ON, the AI may send a longer response as multiple shorter messages.
This can make conversations easier to read and more natural for the customer.
When this option is turned OFF, the AI will usually send replies as a single message.

The End Goal section defines the main outcome the AI Agent should try to achieve during the conversation.
This tells the AI what it should guide the customer toward.
Available goal types may include:
Call to Action
Book Appointment
Send Booking Link

Select Call to Action when you want the AI to guide the customer toward a specific action.
For example, the AI can ask the customer to book a session, visit a link, reply with confirmation, speak with a team member, or complete the next step.
After selecting this option, enter the Call to Action Message.

This message will be used as the final action when the AI reaches the goal.

Select Book Appointment when you want the AI to book an appointment directly on your calendar.
After selecting this option, choose the calendar provider and select the connected calendar.

Calendar Provider: Google Calendar
Select Calendar: Your connected calendar account

This option is useful when the AI should qualify the lead and then schedule a meeting, consultation, demo, or appointment.

Select Send Booking Link when you want the AI to send a booking link to the customer instead of booking the appointment directly.
After selecting this option, choose the calendar provider, select the calendar, and add the booking link text.

Calendar Provider: GHL Calendar
Select Calendar: Your calendar
Booking Link Text: Here is our booking link to schedule a call.
This option is useful when you want the customer to choose a suitable time by themselves.

The Lead Management section controls what should happen if a customer is not qualified or does not want to continue.

The Disqualification Trigger field is used to add words or phrases that mark a lead as lost, unqualified, or not interested.
If the customer sends any of these words or phrases, the AI can understand that the lead should not continue in the normal sales flow.

The Execution Window section controls when the AI Agent is allowed to message contacts.
This is useful if you only want the AI to send messages during business hours.

Select the correct timezone for your business.

Karachi (GMT+5)
Choosing the correct timezone is important because all start and end times follow this timezone.

Set the time window during which the AI Agent is allowed to send messages.

Start: 09:00 AM
End: 05:00 PM
This means the AI Agent will only send messages during this selected time window.

Enable Skip Weekends if you do not want the AI Agent to message contacts on Saturday and Sunday.
This is useful for businesses that operate from Monday to Friday.

The Outside the Window option controls what happens if a message needs to be sent outside the allowed time.

Queue until next window

Do not Send

If this option is selected, messages will wait and send when the next allowed time window starts.
This helps avoid sending messages too early, too late, or outside business hours.

The Also apply to AI replies option controls whether the execution window should also apply to AI replies.
If this option is turned ON, the AI will only reply during the selected business hours.
If this option is turned OFF, some replies may still be sent outside the selected window depending on the setup.
For strict business-hour control, this option should be enabled.

The Knowledge step is where you provide business information to the AI Agent.
This helps the AI understand your company, products, services, pricing, policies, and common customer questions.
The better the knowledge you provide, the better and more accurate the AI responses will be.

The Company Information section is used to give the AI general context about your business.
You can add information such as what your company does, what products or services you offer, your value proposition, target customers, pricing details, business location, delivery or service areas, and important policies.

Our company is a plant seller company in Australia. We provide indoor plants, outdoor plants, low-maintenance plants, and plant care guidance for homes and offices.

This information helps the AI answer general customer questions correctly.

The FAQs section is used to add common questions and answers that the AI can reuse during conversations.
Click Add FAQ to add a new FAQ.
Each FAQ includes:

The customer question or objection.

The response the AI should give.

The AI Agent needs correct information to give accurate answers.

If a customer asks:
What is your price?
The AI should know the correct pricing.
If a customer asks:
Do you deliver to my area?
The AI should know the delivery details.
If a customer asks:
Can I book an appointment?
The AI should know the booking process.
Clear knowledge helps the AI answer faster, more accurately, and more professionally.

The Tasks step is optional inside the AI Agent Builder.
This step is used to define what should happen after the AI reaches its goal or when a specific action is required.
However, for AI Agents, advanced workflow actions should be created from the main Automations page.

AI Agent conversations are created inside the AI Agent Builder, but workflow actions should be created from the Automations page.
After creating the AI Agent, go to:
Automations → Add Workflow
From there, you can build automation flows connected with your AI Agent.
The AI Agent handles the conversation.
The Automations page handles the backend workflow actions.

From the Automations page, you can create workflows for an AI Agent by using the AI Trigger, such as:

• Create contact
• Add tag
• Move contact to pipeline
• Send follow-up message
• Notify team member
• Trigger appointment workflow
• Mark lead as qualified
• Handle disqualified leads
• Pause AI for a contact
• Continue AI after a condition
• Send internal notification

While building an AI Agent, you can save your progress as a draft.
Use Save Draft when the agent is not ready to go live yet.
This is useful when you are still writing the conversation flow, need to add more knowledge, want to review the setup later, are waiting for approval before activation, or want to continue editing later.
Saving your progress regularly helps prevent losing changes.

To edit an existing AI Agent:

1. Go to AI Dashboard.
2. Find the agent you want to update.
3. Click the Edit icon.
4. Update the required sections.
5. Save the changes.
6. Test the agent again.

You can update the agent at any time.
Common updates include changing the tone, updating the AI model, adding new FAQs, changing the booking goal, updating disqualification triggers, improving conversation messages, or changing the contact tag.

Duplicating an AI Agent is useful when you want to create a similar setup quickly

If you already have a Lead Qualification Agent, you can duplicate it and create another version for:
Franchise Leads
Product Enquiries
Customer Support
Plant Sales
Appointment Booking
After duplicating the agent, update the important details such as strategy name, agent role, contact tag, tone, conversation flow, knowledge, and automation workflow.
This saves time and keeps your setup consistent.

The export option allows you to download the AI Agent setup as a JSON file.
This is useful for keeping a backup, sharing a strategy, reusing an agent setup, restoring a previous version, or moving the setup to another workspace.
If import is available, you can upload a JSON file to recreate an agent without building it manually again.

Each AI Agent has a status toggle.
When the toggle is ON, the AI Agent is active.
When the toggle is OFF, the AI Agent is paused.
Turn the agent OFF when you are editing the agent, the campaign is paused, the agent needs review, you want human-only replies, or the agent should not respond temporarily.
This gives you full control over when the AI Agent should operate.

To get the best results from your myCRMSIM AI Agent, follow these recommended practices before activating it.
Use a clear and meaningful Strategy Name so the agent can be easily identified from the dashboard.
Choose the correct Agent Role based on the purpose of the agent, such as Sales Development Representative, Appointment Setter, Customer Support Agent, or Product Advisor.
Select a Tone that matches your business communication style. For sales agents, a persuasive or energetic tone can work well. For support agents, a calm and helpful tone is usually better.
Choose the right AI Model based on your needs. Use lighter models for simple conversations and stronger models for complex or premium workflows.
Keep the first message short, clear, and friendly so the customer understands what the AI can help with.
Use multiple message variations where possible so the AI does not send the same message to every customer.
Add clear and accurate company information in the Knowledge step so the AI can answer customer questions correctly.
Add FAQs for common questions, objections, pricing, delivery, booking, and service details.
Set proper Disqualification Triggers so the AI can understand when a lead is not interested or not qualified.
Configure the Execution Window correctly so the AI only messages contacts during the allowed business hours.
Use the correct End Goal based on your objective. Select Call to Action, Book Appointment, or Send Booking Link depending on what you want the customer to do.
For CRM tracking, always use Add Tag properly. Tags help organize contacts, identify lead status, and trigger the right automation.
Create AI Agent workflows from the main Automations page when advanced actions are needed, such as applying tags, moving leads, sending follow-ups, notifying team members, or managing qualified and disqualified leads.
Always test the AI Agent before turning it ON. Test different customer scenarios, including interested leads, not interested leads, price questions, booking requests, objections, and incomplete replies.
Review the agent after it goes live to make sure replies, tags, automations, and booking actions are working correctly.

Avoid using unclear names like Test Agent or New Bot.
Avoid choosing a role that is too generic.
Avoid writing a very long first message.
Avoid activating the agent without testing it.
Avoid adding incomplete or vague knowledge.
Avoid missing the correct contact tag.
Avoid forgetting to create automation workflows from the Automations page.
Avoid using the wrong end goal for the agent.
Avoid sending messages outside business hours if your business does not support it.
Avoid deleting an important agent without exporting a backup first.

The myCRMSIM AI Agent Builder helps you create powerful AI Agents for sales, support, lead qualification, appointment booking, and CRM automation.
To create an AI Agent, start from the AI Dashboard, click Create New AI Agent, and complete each setup step:
Instructions
Conversation
Configurations
Knowledge
Tasks
The AI Agent should manage the customer conversation, while the Automations page should manage workflow actions such as applying tags, moving leads, sending follow-ups, and notifying team members.
For CRM tracking, the most important action is Add Tag. Tags help organize contacts, identify lead status, and trigger the correct automation.
Before turning the agent ON, always test the conversation, review the knowledge, confirm the end goal, check the execution window, and make sure the correct automation workflow is connected.