Microsoft

Copilot Studio
for Youth
March 18th, 2025
 This session is divided into 4 Modules

• Module 1: Build an Initial Agent with Microsoft

Session Copilot Studio

• Module 2: Manage Topics in Microsoft Copilot Studio
Structure • Module 3: Work with Entities and Variables in
Microsoft Copilot Studio
• Module 4: Enhance Microsoft Copilot Studio Agents
BUILD AN INITIAL
AGENT WITH
MICROSOFT COPILOT
STUDIO

MODULE 1
 Introduction to Microsoft Copilot Studio and AI agents

Steps to create an AI agent easily

Conversational builder for designing interactive conversations

Agenda Generative AI and generative answers explained

Enabling, testing, and optimizing generative answers

Deployment options for Microsoft Copilot Studio agents

Managing live agents and improving performance

Handling user inputs, feedback, and interactions

Future advancements in Copilot Studio AI

Introduction to Microsoft Copilot Studio
Purpose and Function
• Helps users build AI-powered conversational agents
• Provides a no-code, drag-and-drop interface

Key Features
• Supports conversational and generative AI responses
• Allows integration with multiple platforms and channels

Use Cases
• Assists in customer service automation
• Enhances user interaction in business applications
What is an AI Agent?

• AI agent interacts with users through natural language

Definition and Role • Uses machine learning to improve responses over time

• Understands and processes text-based queries

Core Capabilities • Provides automated responses based on predefined
logic

Examples of AI • Chatbots in customer support systems

Agents • Virtual assistants like Microsoft Copilot and ChatGPT
Explore Copilot Studio home
Select the environment in which
to create the copilot

Describe the desired copilot

features and capabilities in a
chat interface

or
Select a template for common
copilot implementations
Create a new copilot in Copilot Studio
• New Copilot
• New Copilot action
• Start with a template
Create a copilot Details dialog
Name and describe the copilot

• Name your copilot

• Select icon
• Describe the purpose of your copilot
• Provide instructions for your copilot
• Select language
• Add knowledge sources
Creating an Agent Using the Conversational
Builder
Overview of Provides a
Uses predefined
graphical interface
Conversational for designing
topics and trigger
phrases
Builder conversations

Building a Define user inputs Use branching logic

Conversation and expected to handle different
responses scenarios
Flow

Testing the Simulate user

interactions to
Adjust settings for
Conversation check responses
better accuracy
Basics of Conversational AI

Understanding Conversational AI

Key Components
AI that processes and
responds to human language Benefits of Conversational AI
Uses Natural Language Intent recognition for
Processing (NLP) to identifying user needs
Context awareness for Enhances user experience
understand queries
meaningful interactions with instant responses
Reduces manual workload in
customer support
Designing Interactive Conversations
Creating Engaging Dialogues

• Use clear and simple language in responses

• Avoid unnecessary complexity in conversation flow

Handling User Queries

• Provide multiple response options for better engagement

• Use fallback messages for unknown inputs

Enhancing User Experience

• Maintain natural conversational tone

• Adjust agent responses based on user feedback
Introduction to Generative AI in Agents
What is Generative AI?
• AI that generates responses based on context
• Uses machine learning to improve interaction

Generative AI in Copilot Studio

• Enhances chatbot flexibility and accuracy
• Provides dynamic responses instead of scripted replies

Advantages of Generative AI
• Reduces manual response creation effort
• Adapts responses based on real-time user queries
What Are Generative Answers?

Definition and • AI-generated responses based on available data

Purpose • Improves chatbot efficiency and user engagement

How Generative • Extracts information from knowledge sources

Answers Work • Uses AI models to create accurate replies

When to Use • Handling complex user queries dynamically

Generative Answers? • Reducing dependency on predefined scripts
Benefits of Generative Answers in AI Agents
Improved Enhanced User Scalability for AI
Response Accuracy Experience Agents
• Provides precise • Adapts responses • Supports handling
answers based on to different a wide range of
existing knowledge conversation queries
• Reduces errors contexts • Reduces manual
from predefined • Offers updates to chatbot
static responses personalized responses
replies for better
engagement
AI general knowledge
Allow AI to use its own general knowledge

• Accessibility: The copilot can

instantly access a vast repository
of information and expertise
across a wide range of subjects

• Versatility: It's capable of

addressing diverse topics and
tasks, making it a versatile
resource for a variety of needs
 Generative answers
The supported knowledge sources are:

Public Website SharePoint Dataverse File Upload

Searches the query input on Connects to a SharePoint Connects to the connected Searches uploaded
Bing, only returns results URL, uses GraphSearch Dataverse environment and documents, returns results
from provided websites to return results uses retrieval-augmented from the document contents
generative technique in
Dataverse to return results
Test panel
Show and hide by selecting Test button

• Ask questions or describe what you

need
• View the results
Publish
• Share with others
• Demo website
• Make the copilot available on channels
How to Deploy an Agent in Different Channels
Understanding AI Agent Performance Metrics
Key Analyzing
Optimizing
Performance Chatbot
Agent Efficiency
Indicators (KPIs) Performance

User engagement Identifying Refining AI

rate and common failure models for better
response points in predictive
accuracy conversation flow responses

Number of Reducing
Adjusting AI logic
successful vs. response time to
based on real-
failed improve user
time usage data
interactions satisfaction
Future Updates and Advancements in Copilot
Studio
Upcoming AI Features

• Improved natural language processing for better understanding

• Advanced analytics for real-time chatbot optimization

Trends in AI Chatbot Development

• Integration with more enterprise tools and applications

• Expansion of AI learning for predictive responses

Preparing for Future Enhancements

• Keep agents updated with new features

• Adapt AI strategies based on industry advancements
 Microsoft Copilot Studio provides tools for building AI-powered agents.

AI agents use conversational AI for real-time user interactions.

The conversational builder helps design smooth and logical conversations.

Summary Generative AI enhances agent responses with contextual understanding.

Generative answers allow AI to provide intelligent automated responses.

Agents can be deployed across multiple channels like Teams and websites.

Managing agents involves refining responses and improving user engagement.

AI agents learn and adapt based on user inputs and feedback.

Future updates will introduce advanced AI features and automation.

https://learn.microsoft.com/th-th/training/modules/create-copilots-copilot-studio
MANAGE
TOPICS IN
MICROSOFT
COPILOT
STUDIO

MODULE 2
 Understanding agent topics and their importance

Creating and organizing topics efficiently

Structuring topics for smooth AI conversations

Agenda Branching topics for handling multiple scenarios

Creating topics from existing support content

Managing topics through editing and refining

Merging, deleting, and archiving topics effectively

Testing topics for accuracy and user experience

Optimizing topics using AI learning and feedback

Topics
• Each topic is a small individual
conversation on a specific
subject

• Multiple topics can be used

together in a single copilot to
provide the customer with an
automated conversation that
feels natural and flows
appropriately
Understanding Agent Topics in Copilot Studio
Definition and Importance
• Topics define how AI agents respond to user queries
• Organizing topics improves chatbot efficiency

Types of Topics
• System topics: Predefined for handling core interactions
• Custom topics: Created to address specific business needs

How Topics Work in Copilot Studio

• Triggered based on user input keywords
• Can be branched for better response handling
Topics in Copilot Studio
• Topics define dialog flows for copilot interaction
• They are initiated by triggers, which can be user input or events such as errors

Greeting Check Weather Goodbye Error

Trigger: Trigger: Trigger: Trigger:
“Hello” “What is the “Perfect! Have a An error occurred
weather today?” great day”
Message 1: Message:
“Hi I’m a Virtual Message: Message: “I’m sorry.
Agent.” “I can help you with “Thanks for Something went
that.” chatting with me. wrong.”
Message 2: Bye!”
“How can I help Question:
you?” “What city do you
live in?”
Creating a New Topic from Scratch

Steps to Create a Topic

Configuring Responses
Click "New Topic" in
Copilot Studio Testing and Refining
Add bot replies for
Define trigger expected user Run simulations to
phrases to initiate inputs ensure smooth
the conversation Use multiple interactions
response options Adjust topic settings
for flexibility based on test
results
Editing and Refining Topics for Better Responses
Why Topic
• Improves chatbot accuracy and efficiency
Refinement is • Ensures smooth user interactions without disruptions
Important

• Modify trigger phrases for better intent detection

Editing a Topic • Adjust bot responses for clarity and relevance

Refining Based on • Analyze past conversations to improve accuracy

User Input • Use AI feedback to enhance chatbot learning
Branching a Topic – What It Means
Real-Life Examples of Branching Topics
• Assisting users in choosing a product
E-commerce Chatbot category
Example • Providing tailored recommendations
based on responses
• Guiding users through
Customer Support Chatbot troubleshooting steps
Example • Offering different solutions based on
issue severity
• Collecting details about service type
Service Appointment and availability
Booking Example • Confirming user preferences before
finalizing booking
Structuring Topics for Easy Navigation
Organizing Topics Logically

Group related topics under relevant categories Ensure a smooth transition between different topics

Avoiding Overcomplication

Keep topic names clear and simple Use intuitive labels for easy identification

Optimizing User Experience

Ensure chatbot follows a natural conversation flow Minimize the number of steps to reach a resolution
Merging Similar Topics to Avoid Redundancy
Identifying Duplicate Topics
• Review topics with overlapping responses
• Analyze user interactions to find repetitive queries

Steps to Merge Topics

• Combine similar topics into a single structured response
• Adjust trigger phrases to cover all relevant queries

Benefits of Merging Topics

• Reduces complexity in chatbot management
• Improves response accuracy and efficiency
Deleting and Archiving Unused Topics
• If a topic is outdated or no longer
relevant
When to Remove Topics • If user interactions show low
engagement
• Archive topics to retain data for
future reference
Archiving vs. Deleting • Delete topics permanently if they are
no longer needed
• Reduces clutter in chatbot
Impact on Chatbot management
Efficiency • Ensures a streamlined and optimized
conversation flow
 Topics help structure AI conversations for better user engagement.

Well-organized topics improve AI agent accuracy and response

quality.
Topic branching allows AI to handle multiple user scenarios.

Summary Support content can be converted into AI-driven topics.

Editing and refining topics enhances the agent’s efficiency.

Unused topics should be merged, deleted, or archived for clarity.

Testing ensures topics function correctly within conversations.

AI learning and user feedback improve topic performance over time.

Optimizing topics creates a better and more interactive AI

experience.
Celebrate your accomplishments and feel
empowered
• Go to Achievements on your Microsoft Learn
profile
• Redeem the code provided by your trainer
• Find your new badge on Modules, Courses &
More
• Share your new achievement with your
professional network. You can also download
and print your certificate.
WORK WITH
ENTITIES AND
VARIABLES IN
MICROSOFT
COPILOT STUDIO

MODULE 3
 Understanding entities and their role in AI conversations

Using prebuilt and custom entities effectively

Creating, managing, and editing custom entities

Agenda Using entities to improve AI agent responses

Variables and their role in storing information

Types of variables and their practical applications

Reusing variables across different AI topics

Managing structured data with entities and variables

Enhancing AI interactions through personalized responses

Introduction to Entities in Copilot Studio

Types of Entities
• Prebuilt entities for common data
recognition
• Custom entities for domain-specific
input processing
Where Entities Are Used
• Extracting user-provided details
from conversations
• Improving chatbot logic with
structured data
Prebuilt entities
Commonly used
information
•Person
•Location
•Numbers
•Date and times
•Colors
Create a custom entity
Entity types

•Closed list
•Regular
expression
Closed list entity
List of items

•Add names
•Add synonyms
Regular expression entity
Regex

•Specify a regex pattern

•.NET regular expressions syntax
•Use cases
•Find specific character patterns
•Validate text to ensure that it matches a predefined
pattern
•Extract, edit, replace, or delete text substrings
Real-Life Examples of Entities in Chatbots
E-commerce Healthcare Banking
Chatbot Chatbot Chatbot

Extracts transaction
Recognizes product
Identifies symptoms details like account
names and order
from user input numbers and
numbers
payment amounts

Suggests relevant Provides accurate

Assists in tracking
health information responses related to
orders and
based on recognized financial
processing returns
data transactions
Create variables
Create in Question node

• Created automatically
• Rename
Use variables
{x} icon

• Insert variable into a node

Different Types of Variables in AI Agents
• Automatically generated by Copilot Studio
System Variables • Used for managing session details and user
information

• Created to store specific user-provided data

Custom Variables • Can be used across different topics in conversations

• Session variables exist only within a single

Session-Based vs. conversation
Global Variables • Global variables persist across multiple interactions
How to Reuse Variables Across Topics
Benefits of Reusing Variables

• Reduces repetitive data entry for users

• Improves chatbot efficiency in long conversations

Methods for Variable Reuse

• Share stored values across multiple chatbot topics

• Retrieve previous user inputs to maintain context

Avoiding Variable Overwrites

• Ensure correct naming conventions for variables

• Use clear logic to prevent conflicts in data retrieval
Best Practices for Organizing Variables
Preventing Maintaining a
Defining Clear
Unnecessary Logical Variable
Variable Structures
Storage Framework
• Use descriptive • Limit data • Align variable
names for easy retention to usage with chatbot
identification essential details workflow
• Categorize • Remove expired • Avoid duplication
variables based on variables to by reusing stored
conversation optimize chatbot values
needs performance
Using Entities and Variables Together
• Entities extract structured data from
How Entities and user inputs
Variables Work Together • Variables store extracted information
for future use
• AI agent captures dates using an
entity and stores them in a variable
Practical Use Cases • User preferences are extracted and
reused across different topics
• Improves chatbot memory and
Benefits of Combining personalization
Both • Enhances efficiency in handling
complex conversations
Advanced Use of Entities and Variables
Scenario-Based Optimizing AI
Dynamic Data
Entity and Accuracy with
Handling in AI
Variable Usage Data Storage

Travel chatbot:
AI agents retrieve Ensures relevant
Extracts destination
stored variables for data is retained
from entity and
personalized across chatbot
stores date in a
interactions interactions
variable

HR chatbot:
Entities ensure Enhances response
Captures job role
structured efficiency by
using an entity and
recognition of user- minimizing repeated
retains applicant
provided information user input
details
 Entities help AI agents recognize and process user inputs correctly.

Prebuilt entities simplify common data recognition tasks.

Custom entities allow agents to handle unique business needs.

Entities enhance agent responses by structuring input data.

Summary
Variables store and manage dynamic conversation information.

Different variable types enable AI agents to track user details.

Reusing variables improves conversation flow across multiple

topics.
Structured data enhances personalization and agent intelligence.

Entities and variables together create smarter, more efficient AI

interactions.
https://learn.microsoft.com/th-th/training/modules/create-copilots-copilot-studio
ENHANCE
MICROSOFT
COPILOT STUDIO
AGENTS

MODULE 4
 Enhancing AI agents using Power Automate

Connecting Power Automate for automated actions

Understanding different trigger types in AI agents

Agenda Setting up and managing automated triggers

Analyzing agent performance using AI insights

Measuring user engagement and response accuracy

Troubleshooting AI agents for better efficiency

Optimizing AI agents based on user feedback

Future advancements and AI-driven improvements

Enhancing AI Agents with Power Automate
• A Microsoft tool for automating workflows
What is Power
Automate? • Connects AI agents to external applications and
services
How Power • Automates repetitive tasks like data retrieval
Automate
Enhances AI • Improves chatbot efficiency with seamless
Agents integrations

• Automating ticket creation for customer support

Common Use
Cases • Triggering notifications based on chatbot
interactions
Introduction to Power Automate for AI Agents
Key Features of Power Automate

No-code interface for building automation flows Supports API connections for advanced automation

Role in AI Agent Optimization

Reduces manual intervention in chatbot workflows Ensures faster and more accurate response generation

Supported AI Integrations
Connects with Microsoft Teams, Outlook, and Dataverse Supports third-party applications via API connectors
Use Power Automate to add actions
Access information from
other systems

• Actions tab
• Add an action
• Create a new flow
Adding Automated Actions to Agents

Types of Automated • Sending notifications through email or Teams

Actions • Fetching and displaying real-time data to users

Steps to Add an • Define trigger conditions in Power Automate

Automated Action • Configure actions to execute based on user inputs

Benefits of
• Reduces chatbot response time for user queries
Automating AI • Enhances interaction with real-time data processing
Responses
What Are Trigger Types in AI Agents?
• Conditions that initiate specific
chatbot actions
Definition of Triggers • Used to automate responses based
on user inputs
• Keyword-based triggers: Activated by
Types of Triggers in AI specific user phrases
Agents • Event-based triggers: Respond to
system or external events
• Ensures chatbot responds
Importance of Using appropriately to different scenarios
Triggers • Reduces manual intervention by
automating interactions
How to Choose the Right Trigger Type
Factors to Consider
Comparison of Optimizing Trigger
When Selecting a
Trigger Types Selection
Trigger
• Nature of chatbot • Keyword triggers: • Ensure triggers align
interaction (static Best for FAQ-style with business logic
vs. dynamic chatbot interactions and user needs
responses) • Event-based • Avoid overloading
• Need for real-time triggers: Ideal for chatbot with
updates or real-time process unnecessary
scheduled automation triggers
automation
Event-Based Triggers in AI Agents
What Are Examples of Advantages of
Event-Based Event-Based Event-Based
Triggers? Triggers Triggers

Automated Sending Enables

responses confirmation seamless
activated by emails after user integration with
external events registration external systems

Commonly used Updating Reduces

for real-time database entries dependency on
notifications and based on chatbot user-initiated
workflow actions interactions inputs
How to Set Up and Test Triggers

Optimizing Trigger
Performance
• Refine trigger conditions
Testing the Trigger for based on chatbot usage
Accuracy analytics
• Ensure triggers do not
• Simulate real-world user
Steps to Create a conflict with other
inputs to verify
automation flows
Trigger in Copilot functionality
Studio • Monitor chatbot logs to
detect misfired triggers
• Define conditions that
activate the trigger
• Configure corresponding
chatbot responses or
actions
Understanding AI Agent Performance Metrics
Why Performance Metrics Matter

• Helps assess chatbot effectiveness in real-world use

• Identifies areas for improvement in AI responses

Key AI Agent Metrics

• Response accuracy: Measures correctness of chatbot answers

• Engagement rate: Tracks user interactions with the chatbot

Optimizing AI Agent Performance

• Regularly analyze chatbot logs for inconsistencies

• Adjust conversation flows based on performance data
How to Analyze Agent Performance in Copilot
Studio
Accessing • Use built-in analytics tools in Copilot Studio
Performance Reports • Monitor chatbot response times and success rates

Identifying
• Detect common user drop-off points
Performance • Assess errors in chatbot-triggered actions
Bottlenecks

Improving AI Agent • Refine conversation paths for better engagement

Efficiency • Optimize chatbot responses based on user behavior
Measuring User Engagement and Satisfaction
Improving
Importance of
Ways to Measure Engagement
Engagement
Engagement Through AI
Metrics
Optimization

Determines Tracking session Personalize

chatbot’s duration and responses to
effectiveness in message enhance user
user interactions exchange count experience

Analyzing user Reduce chatbot

Helps in refining
satisfaction response time for
AI conversation
ratings and better interaction
design
feedback flow
Using AI Analytics to Improve Agent Responses
Role of AI Analytics in Chatbots
• Provides insights into chatbot response accuracy
• Identifies areas for improvement based on user interactions

Key AI Analytics Tools in Copilot Studio

• Conversation logs for tracking user queries
• Performance dashboards to monitor chatbot success rates

Optimizing AI Responses Using Data

• Adjust chatbot replies based on frequent user concerns
• Use sentiment analysis to improve response tone
Troubleshooting AI Agents for Errors
Common Chatbot Issues
Looping errors where chatbot repeats responses
Misinterpreting user input due to incorrect trigger phrases
unnecessarily

Steps for Troubleshooting AI Agents

Review chatbot logs to identify error patterns Adjust entity recognition for improved input processing

Ensuring Chatbot Stability

Regularly test different chatbot scenarios Use fallback messages for handling unexpected inputs
Optimizing AI Agents for Better Conversations
• Ensure smooth transitions between different chatbot topics
Improving Chatbot
• Minimize unnecessary user prompts for a streamlined
Flow experience

• Train chatbot with diverse user queries for better recognition

Reducing Chatbot
• Improve NLP (Natural Language Processing) to enhance
Confusion understanding

Personalizing
• Use stored user data to provide relevant responses
Conversations for • Adjust chatbot tone based on user interaction history
Better Engagement
Best Practices for AI Agent Performance
Improvement
Ensuring Chatbot Improving AI Enhancing Overall
Reliability Training Methods User Experience

• Regularly update AI • Use real-world data • Optimize chatbot

models for better to refine chatbot speed for faster
accuracy responses interactions
• Monitor chatbot • Incorporate • Provide multiple
logs for unexpected feedback loops for response options to
behavior continuous learning improve user
engagement
Using AI Insights to Enhance User Experience
Role of AI Insights in Chatbot Optimization
• Helps refine chatbot responses based on user behavior
• Identifies patterns in conversation flow for improvement

Key AI Insights in Copilot Studio

• Tracks user satisfaction through feedback analysis
• Detects conversation drop-off points for optimization

Enhancing User Experience with AI Insights

• Adjust chatbot responses based on frequently asked questions
• Personalize interactions by analyzing past conversations
How AI Agents Adapt to User Needs Over Time
Improving AI
AI Learning and Methods for AI
Agent
Adaptation Adaptation
Flexibility

Adjust chatbot
Uses machine
Regularly update behavior based on
learning to refine
chatbot training data real-time
chatbot interactions
interactions

Continuously Optimize NLP

Use feedback loops
improves responses models to enhance
to fine-tune chatbot
based on user contextual
responses
engagement understanding
Improving AI Accuracy with User Feedback
Importance of User • Helps in identifying chatbot response gaps
Feedback in AI • Allows real-time improvements in AI-generated
Accuracy answers

Methods to Collect • Use rating systems after chatbot interactions

Feedback • Analyze missed responses for AI model refinements

• Update chatbot training data based on user inputs

Refining AI Models for
• Optimize chatbot triggers to reduce incorrect
Better Accuracy responses
Managing AI Agent Updates and New Features
Why Regular • Keeps AI agent aligned with business needs and user
Updates Are expectations
Important • Ensures chatbot is using the latest AI advancements

• Test all updates before deploying in a live

Best Practices environment
for AI Updates • Monitor chatbot performance after implementing
new features
Introducing New • Implement AI-driven automation for repetitive tasks
Features to
Improve AI • Enhance chatbot personalization with advanced
Efficiency analytics
Handling AI Agent Compliance and Security
Importance of Compliance in AI Chatbots
• Ensures AI agents follow industry regulations and policies
• Protects user data from unauthorized access

Key Security Considerations

• Implement encryption for sensitive user information
• Use role-based access control to manage permissions

Ensuring AI Compliance
• Regular audits to check chatbot data privacy standards
• Align chatbot policies with regulatory frameworks like GDPR
Preparing for the Future of AI Conversations
Emerging Future-
Advancements
Trends in AI Proofing AI
in AI Learning
Chatbots Agents

Improved natural
AI adapting based Regular updates to
language
on user behavior keep pace with
understanding for
and preferences evolving technology
better interactions

Integration with Automated training

Adopting ethical AI
multimodal AI for models for
practices to
voice and text continuous chatbot
enhance user trust
responses improvement
Next Steps for Using Microsoft Copilot Studio
Ongoing AI
Maximizing Copilot Expanding AI Agent
Optimization
Studio Capabilities Functionality
Strategies
• Explore advanced • Implement more • Regularly review
integrations with automation through chatbot
Microsoft services Power Automate performance
• Use analytics to • Enhance chatbot metrics
continuously refine customization with • Align chatbot
chatbot dynamic data updates with
performance processing evolving business
needs
 Power Automate helps AI agents perform automated actions seamlessly.

Integrating automation improves AI efficiency and response speed.

Triggers define how AI agents react to specific user actions.

Summary Choosing the right trigger ensures accurate agent responses.

Performance analysis helps in optimizing AI agent interactions.

User engagement metrics guide improvements in agent design.

Troubleshooting helps resolve errors in AI agent conversations.

Optimized AI agents provide better and more accurate responses.

AI advancements continue to enhance agent functionality and automation.

Celebrate your accomplishments and feel
empowered
• Go to Achievements on your Microsoft Learn
profile
• Redeem the code provided by your trainer
• Find your new badge on Modules, Courses &
More
• Share your new achievement with your
professional network. You can also download
and print your certificate.
Thank you for
listening!
Feel free to send in any questions
to [email protected].