Explore how Trasol built an internal AI assistant using Microsoft Copilot Studio to help UAE consultants with regulations, pricing, jurisdiction recommendations, document drafting and onboarding
Business consultants often need to find information quickly while working with clients. Questions around jurisdictions, business packages, pricing, UAE regulations, compliance deadlines, documentation and client objections can require checking multiple documents or asking experienced team members.
For one such business environment, we built an internal AI assistant using Microsoft Copilot Studio.
The solution, named Zone Bot, was designed to give consultants and staff a single conversational interface for finding business information and handling common queries.
Instead of searching through multiple documents manually, employees can ask the chatbot a question in natural language and receive an answer based on the configured business knowledge and conversation topics.
6
| Area | Details |
|---|---|
| Project Type | Internal AI Assistant |
| Platform | Microsoft Copilot Studio |
| Agent Name | Zone Bot |
| Primary Users | Consultants and internal staff |
| Environment | Microsoft Power Platform |
| Primary Channel | Microsoft Teams / Microsoft 365 |
| Knowledge | Business documents and internal guidance |
| Automation | Topics, knowledge retrieval and conversational workflows |
| Use Case | Business setup and consulting support |
The chatbot was created as an internal assistant for Creative Zone UAE, supporting consultants with business setup queries, jurisdiction recommendations, package pricing, client queries, UAE regulations, document generation and staff onboarding.
The screenshots provided for this case study show the actual Copilot Studio configuration and components used in the solution.
The consulting team had to deal with information spread across different documents and processes.
Typical questions could include:
Traditionally, answering these questions could involve:
Question → Search documents → Check multiple files → Ask another employee → Confirm information → Respond to client
This created unnecessary dependency on experienced employees and made it harder to provide consistent answers.
The goal was to introduce a conversational AI layer over the organization's existing knowledge.
We designed and configured Zone Bot using Microsoft Copilot Studio as an internal AI assistant.
The solution combines:
Copilot Studio allows organizations to add knowledge sources to agents so responses can be grounded in enterprise information. Knowledge can be added at the agent or topic level.
The first screenshot shows the Zone Bot overview and configuration.
The agent contains a business-specific description explaining its purpose:
Internal AI assistant for Creative Zone UAE.
The configured assistant is intended to help with:
The agent also has dedicated instructions defining how it should behave and respond to users.
This is important because an enterprise chatbot should not simply behave like a generic AI assistant.
Its responses need to follow the organization's business context and instructions.
One of the important parts of the solution is the Knowledge section.
The screenshot shows four business documents connected to Zone Bot:
Used to provide information related to business processes and expected timelines.
Used when consultants need information about packages and pricing.
Provides guidance for handling common client objections.
Provides information used when discussing jurisdiction-related questions.
These knowledge sources allow the assistant to answer questions based on the organization's configured information rather than relying only on general AI knowledge.
Copilot Studio supports knowledge sources from enterprise data and can use them to ground agent responses.
The chatbot also contains multiple custom topics.
The screenshot shows topics such as:
These topics provide a structured way of handling specific types of conversations.
For example:
A consultant can ask about the cost of a particular package and the bot can use the configured pricing knowledge.
The assistant can help consultants find relevant jurisdiction information based on the client's requirements.
The bot can assist with creating or structuring business documents based on the configured process.
The assistant can help users understand compliance-related dates and requirements based on the available business information.
New employees can use the assistant to find information about internal processes and procedures.
Copilot Studio supports custom topics alongside knowledge-based responses, allowing organizations to combine conversational guidance with enterprise information.
The objective was not to make employees remember commands.
Instead, users can communicate with the assistant naturally.
For example:
Consultant:
Which jurisdiction should I consider for this client?
Zone Bot:
Based on the configured jurisdiction guidance, here are the relevant options and considerations...
Another example:
Consultant:
What is the cost of this package?
Zone Bot:
Based on the current pricing guide, the applicable package information is...
This conversational approach makes the internal knowledge base much easier to access.
The agent instructions define the role and expected behavior of Zone Bot.
The configured instructions establish that the chatbot is an internal assistant for Creative Zone UAE and that it supports areas such as:
This layer is particularly important in enterprise AI projects.
A generic AI model can provide broad information, but an enterprise assistant needs to understand:
Who it is → Who it serves → What information it should use → What tasks it should support → How it should respond
The chatbot was configured for Microsoft channels, including Microsoft Teams and Microsoft 365.
This is important because employees do not need to open a separate application just to interact with the assistant.
The Copilot Studio publishing model allows an agent to be connected to channels after publishing, including Microsoft Teams and Microsoft 365 Copilot.
For an internal organization, this provides a practical experience:
Employee → Microsoft Teams → Zone Bot → Business Knowledge / Topics → Response
The agent can also be distributed through the organization's Teams environment, depending on the organization's configuration and approval process.
The screenshots also show an important configuration warning:
Your agent does not require end user authentication.
This means the agent was configured at that point to allow anonymous users to interact with it.
For an internal enterprise chatbot, authentication should be carefully considered because the assistant may expose internal business information.
Microsoft documents that Copilot Studio agents can use Microsoft authentication and that authentication can be used to restrict access to specific users or groups.
For a production internal implementation, the recommended architecture would therefore be:
Employee → Microsoft Entra ID → Teams/Microsoft 365 → Copilot Studio Agent → Authorized Knowledge
This provides a stronger access model than anonymous access.
The screenshot also shows another configuration warning:
Your agent will search the web for answers.
This is an important consideration for an enterprise chatbot.
Web search can be useful when the organization intentionally wants current external information. However, business-critical answers should generally be grounded in approved internal knowledge where appropriate.
For example:
Pricing, internal processes, objection handling and company procedures should come from approved business sources.
Public regulatory or general industry information may require web access, subject to the organization's governance requirements.
This separation can help reduce the risk of the chatbot using outdated or irrelevant information.
The configuration also includes a Suggested Prompts section.
Suggested prompts can make the assistant easier for employees to use because users immediately see examples of what they can ask.
For example:
This is particularly useful when employees are using an AI assistant for the first time.
The Zone Bot configuration includes monitoring information such as:
This provides a starting point for understanding how employees interact with the assistant.
For a production implementation, we would extend this monitoring approach to track areas such as:
This creates a continuous improvement cycle:
User Question → Bot Response → Monitor → Identify Gap → Update Knowledge/Topic → Test → Publish
| Module | Purpose |
|---|---|
| AI Assistant | Central conversational interface |
| Knowledge Base | Provides business-specific information |
| Pricing Guidance | Supports package and pricing queries |
| Jurisdiction Guide | Supports jurisdiction-related questions |
| Process & Timeline | Provides process information |
| Objection Handling | Helps consultants respond to common objections |
| Document Drafter | Supports document preparation |
| Compliance Tracker | Supports compliance-related questions |
| Staff Onboarding | Provides internal onboarding assistance |
| Arabic-English Support | Helps bridge language requirements |
| Meeting Preparation | Supports consultant preparation |
| License Activity | Supports license-related queries |
| Monitoring | Tracks engagement and agent performance |
| Teams Integration | Makes the assistant available where employees work |
The goal is not simply to replace document search.
The larger objective is to make organizational knowledge accessible through a conversational interface.
Copilot Studio provides several capabilities that fit this type of internal AI assistant:
Agents can be published to multiple channels, and changes to the agent need to be published before the updated version becomes available to users.
Our implementation approach can be divided into the following phases.
We identified the major categories of questions employees handle regularly.
Business documents were reviewed and organized into relevant knowledge categories.
We configured:
Specific business scenarios were converted into custom conversation topics.
The agent was tested against representative questions to verify whether it could provide appropriate responses.
After testing, the agent was published and configured for the required Microsoft channels.
Usage and conversation behavior can be reviewed to identify new topics, missing knowledge and areas requiring improvement.
The implementation provides a centralized conversational interface for internal business knowledge.
The key improvements include:
Rather than creating another standalone application, the solution brings AI assistance into the employees' existing Microsoft 365 environment.
The current architecture can be extended beyond question-and-answer scenarios.
Potential enhancements include:
The chatbot can be extended to trigger business workflows where appropriate.
For example:
User → Zone Bot → Power Automate → Business System → Result
Connect the assistant to an authorized business system so consultants can retrieve application status.
Connect compliance information with automated reminders and escalation workflows.
Expand document-drafting capabilities and connect them with approved templates.
Provide different information depending on the employee's role and permissions.
Connect approved sources such as SharePoint, Dataverse or other business systems through appropriate connectors and authentication. Copilot Studio supports connections to enterprise services including SharePoint, Dataverse and external APIs.
For questions that require human intervention, the assistant can be designed to route the conversation or provide the appropriate escalation path.
One of the important lessons from this project is that building an enterprise chatbot is not just about connecting an AI model.
The quality of the solution depends heavily on:
Good Knowledge + Clear Instructions + Well-designed Topics + Security + Testing + Continuous Improvement
A chatbot can have a powerful AI model, but if the underlying business information is incomplete or poorly organized, the resulting experience will also be limited.
That is why our approach focuses on the complete solution rather than only the chatbot interface.
Platform: Microsoft Power Platform
AI Platform: Microsoft Copilot Studio
Knowledge: Business documents / enterprise knowledge
Authentication: Microsoft identity and access controls
Channels: Microsoft Teams / Microsoft 365
Automation: Power Automate where required
Data Integration: SharePoint, Dataverse and APIs where required
The Zone Bot project demonstrates how Microsoft Copilot Studio can be used to build an internal AI assistant around an organization's existing business knowledge.
Instead of employees manually searching through pricing documents, process guides, jurisdiction information and other internal resources, they can interact with a single conversational assistant.
The solution combines AI, enterprise knowledge, custom topics, Microsoft 365 integration and business-specific instructions into one internal assistant.
For organizations already using Microsoft 365 and Power Platform, this approach can provide a practical foundation for gradually introducing AI into everyday business processes.
A Copilot Studio chatbot, now generally referred to as an agent, is a conversational AI solution that can use instructions, knowledge, topics, tools and integrations to interact with users.
Yes. Copilot Studio agents can be published and connected to Microsoft Teams and Microsoft 365.
Yes. Knowledge sources can be added to agents so responses can be grounded in enterprise information.
Yes. Agents can use Microsoft authentication, and tools can also be configured to require user authentication where access needs to be restricted.
Yes. Copilot Studio can use connections, tools and APIs to interact with services such as SharePoint, Dataverse and external systems.
Yes. The agent can be configured with business-specific instructions, knowledge, topics, workflows and tools.
At Trasol Technologies, we help organizations design and implement Microsoft Power Platform and AI solutions around their actual business processes.
Our Copilot Studio services can include:
If your organization has large amounts of internal information spread across documents, SharePoint, business applications and processes, a custom Copilot Studio agent can provide a more conversational way for employees to access that information.
Looking to build an internal AI assistant with Microsoft Copilot Studio? Talk to Trasol Technologies about your requirements.