Business process automation is entering a new stage. It is no longer only about designing workflows, executing tasks, and measuring results from a BPM platform. Now, it is also possible to interact with your processes using natural language, directly from the AI assistant you already use to work.
With Flokzu’s new MCP server, you can connect MCP-compatible AI clients such as Claude Desktop, Claude Code, Cursor, VS Code, and other agents that support MCP over HTTP to interact with your Flokzu account in a more direct, flexible, and natural way.
This means you can ask your AI assistant to perform actions inside Flokzu, such as starting process instances, checking records, modifying information, managing users, assigning roles, or exporting process history, always within the permissions defined by your API Key and configuration.
For teams already using Flokzu, this integration opens a new way to operate: fewer manual steps, more agility, and an experience that feels closer to the way people naturally express what they need.
What Is MCP and Why Does It Matter for Process Automation?
MCP stands for Model Context Protocol. It is a protocol that allows AI assistants to connect with external tools, systems, and data sources in a structured way.
In simple terms, MCP works as a bridge between an AI assistant and an application. In this case, Flokzu’s MCP server allows a compatible AI assistant to interact with your Flokzu account through natural language instructions.
This matters because AI assistants are no longer limited to conversation or text generation. With MCP, they can become an intelligent interface for executing actions, retrieving information, and operating real business systems.
For a BPM platform like Flokzu, this connection makes a lot of sense. Business processes involve tasks, data, approvals, users, roles, instances, documents, and history. Being able to interact with all of that from an AI assistant can help teams move faster and access process information more easily.
What You Can Do When You Connect Your AI Assistant to Flokzu
Flokzu’s MCP server allows your AI assistant to interact with different elements of the platform. Once the connection is configured, you can use natural language to request specific actions.
For example, you can ask your assistant to create a new process instance, retrieve data from an existing instance, update field values, list database records, invite users, assign roles, or export process history.
Some available actions include:
- Starting a new process instance.
- Getting data from an existing process instance.
- Updating field values in an existing process instance.
- Listing database records.
- Getting specific records by ID.
- Creating new database records.
- Modifying existing records.
- Inviting new users.
- Removing users from a tenant.
- Listing roles assigned to a user.
- Assigning or removing roles.
- Exporting the history of an instance.
- Exporting the full history with field details.
- Accessing metadata from attached documents.
Instead of manually navigating through different sections, users can write an instruction such as: “Create an instance of the Purchase Request process with an amount of 5,000 USD and ACME as the supplier.” The assistant interprets the request, invokes the corresponding tool, and returns the result.
A New Way to Interact With Your Processes
The connection between Flokzu and AI assistants does not replace process management. It enhances it.
Until now, process automation helped organizations structure operations, define responsibilities, establish rules, generate traceability, and measure performance. With MCP, a conversational layer is added, making it easier to interact with that operation.
This can be especially useful for users who need to quickly retrieve information, start processes without browsing forms, check statuses, execute repetitive actions, or work from tools where they already use AI every day.
For example, an operations manager could ask:
“Show me the latest purchase requests pending approval.”
An administration manager could write:
“Create a new record for this supplier with the following information.”
An IT leader could request:
“Invite this user and assign the corresponding role.”
A compliance team could ask:
“Export the history of this instance for internal review.”
The value is in reducing friction. People can interact with processes, data, and users in Flokzu in a more natural way, without losing structure or control.
Why This Matters for Companies Managing Complex Processes
In medium and large organizations, processes often involve multiple departments, systems, validations, and approval levels. That complexity frequently creates delays, errors, rework, and dependency on technical teams for operational tasks.
Flokzu already enables companies to automate complex processes in a simple way, using forms, BPMN workflows, rules, integrations, metrics, and traceability. The MCP connection adds a new possibility: interacting with that automation from an AI assistant.
This matters because many operational tasks do not necessarily require navigating a full interface. Sometimes, the user only needs to execute a specific action or retrieve specific information.
With MCP, that interaction can become faster. The user expresses what they need in natural language, and the assistant invokes the right tool.
Use Cases for Flokzu’s MCP Server
Creating Process Instances From the Assistant
One of the most direct use cases is starting Flokzu processes from an AI assistant.
This can apply to purchase requests, internal approvals, complaints, supplier onboarding, HR requests, administrative tickets, document management, or any workflow already modeled in Flokzu.
For example:
“Create a new instance of the Supplier Onboarding process with this information.”
“Generate a purchase request for the marketing department for 3,000 USD.”
“Open an internal complaint case with high priority.”
This type of interaction helps reduce manual steps and makes it easier to start processes faster.
Checking Information From Existing Processes
Another important use case is retrieving information from ongoing or completed process instances.
A user can ask the assistant for information about a specific case, the status of a request, the fields associated with an instance, or the history of a process.
This makes it easier to answer questions such as:
“What is the status of this request?”
“What information is loaded in this instance?”
“Who participated in this process?”
“What changes were made in this case?”
For operations, finance, compliance, and support teams, this capability can save time and improve visibility.
Updating Fields in an Instance
The MCP server also allows users to modify field values in an existing instance, always according to the configured permissions.
This can be useful when a team needs to update operational information, correct data, complete pending fields, or register changes without manually entering the platform.
For example:
“Update the approved amount of this request to 7,500 USD.”
“Add this comment to the case.”
“Modify the selected supplier in this instance.”
As in any process automation initiative, control is essential. These actions should always be aligned with the API Key scope, defined permissions, and each organization’s governance rules.
Working With Flokzu Databases
Databases in Flokzu allow teams to centralize information that can later be used in processes. With MCP, the AI assistant can interact with those records.
It can list records, apply filters, retrieve specific data, create new records, or modify existing information.
Some examples include:
“List the latest 20 products with stock below 10.”
“Create a new record for this customer.”
“Update the status of this supplier.”
“Find the record associated with this ID.”
This makes daily operations easier, especially when teams work with data that must be consulted or updated frequently.
Managing Users and Roles
User management can also be assisted from MCP-compatible clients.
The server allows you to invite users, remove users from the tenant, list roles, assign roles, and remove roles.
For example:
“Invite john@company.com as a user and assign him the Approver role.”
“Show me which roles this user has.”
“Remove the Requester role from this user.”
For administration or IT teams, this can help speed up recurring tasks and maintain a more organized access management process.
Exporting History and Evidence
One of the most relevant capabilities for organizations that require traceability is history export.
With MCP, it is possible to request the history of an instance, the full history with field details, and metadata from attached documents.
This can be especially useful for audits, internal reviews, process control, case analysis, or compliance support.
For example:
“Export the full history of this instance.”
“Get the metadata from the attached documents in this process.”
“I need to review the change history of this case.”
In critical processes, having available and traceable evidence is essential.
Security: A Key Part of the Integration
When connecting AI assistants with business systems, security is a natural concern. That is why the connection to Flokzu’s MCP server is designed to operate with clear controls.
To configure the connection, you need an API Key and the email of the administrator user on whose behalf operations will be executed. The scope of the API Key defines which actions the assistant can perform.
This is important because AI should not have unlimited access to the operation. Capabilities must be limited by credentials, permissions, and configuration.
The API Key should also be treated like a password. It should not be published in public repositories or shared through insecure channels.
According to Flokzu’s documentation, the MCP server uses encrypted HTTPS transport, does not persist API Keys, maintains tenant isolation, and records operations in Flokzu’s history associated with the configured email.
This allows organizations to combine agility with control, which is essential for companies automating sensitive processes.
What Data Goes to the AI Provider?
A common question is what information reaches the AI assistant provider.
Flokzu’s documentation explains that only the information the user shows in the chat goes to the AI provider: the inputs the user writes and the outputs returned by the model.
Calls to the MCP server go directly from the local client to Flokzu, without passing through the AI provider’s servers.
This point is especially important for IT, security, and compliance teams, because it helps them better understand the flow of information and define internal usage policies.
Even so, each organization should establish its own rules regarding what data can be shared with AI assistants, which users can use these connections, and what types of actions are allowed.
Best Practices for Using Flokzu’s MCP Server
To make the most of this integration in a secure and organized way, it is a good idea to define best practices from the start.
Define the Right API Key Scope
The API Key determines which actions the assistant can execute. The recommended approach is to apply the principle of least privilege: grant only the permissions required for the intended use case.
Use the Right Accounts and Emails
Operations are associated with the configured email. That is why it is important to define under which user actions will be executed and how they will be audited.
Do Not Share Credentials Through Insecure Channels
The API Key must be treated as a sensitive credential. It should not be shared through open chats, public documents, repositories, or unprotected channels.
Start With Controlled Use Cases
Before extending usage across teams, it is best to test the connection with specific processes and validate that actions behave as expected.
Train Users
Even though the interaction happens through natural language, users should understand what they can ask the assistant, what information they are sharing, and which actions can impact Flokzu.
Review Audit Trails and History
Every operation should be reviewable. This allows teams to control usage, detect errors, and maintain visibility over executed actions.
Align the Integration With Internal AI Policies
Connecting AI assistants to business systems should be part of a broader strategy for responsible AI adoption across the organization.
Which Clients Can You Use to Connect to Flokzu’s MCP Server?
Flokzu’s MCP server can be used with compatible clients that support MCP over HTTP. The documentation mentions options such as Claude Desktop, Claude Code, Cursor, VS Code, and other agents that support MCP over HTTP.
The setup follows a common structure: the MCP server endpoint and authentication headers with the API Key and the Flokzu user email.
This allows different teams to choose the AI client that best fits their way of working, while keeping Flokzu as the process execution layer.
MCP, BPM, and AI: A Combination to Work Better
Process automation with BPM already helps organizations reduce errors, eliminate repetitive tasks, structure approvals, and gain visibility. AI adds a new productivity layer, helping users interpret instructions, retrieve information, and execute actions using natural language.
When both capabilities are connected through MCP, the result is a smoother experience between people and processes.
This does not mean losing control or delegating critical decisions without supervision. It means creating new ways to interact so teams can work with greater agility while maintaining rules, permissions, traceability, and auditability.
In that sense, Flokzu’s MCP server reflects a central idea: AI does not replace process management. It makes it more accessible, faster, and closer to the way people naturally express what they need.
Why This Integration Strengthens Flokzu’s No-Code Approach
Flokzu was built around a clear idea: making process automation simpler, more accessible, and more powerful for organizations. Its no-code approach allows teams to design and improve processes without relying on long development cycles.
The connection with AI assistants through MCP follows the same direction. It allows users to interact with business processes using natural language, reducing technical barriers and accelerating operational tasks.
For business teams, this means more autonomy.
For IT, it means a controlled way to enable new capabilities without building custom integrations from scratch for every interaction.
For operations, it means more speed and less friction.
For compliance, it means maintaining history and traceability over executed actions.
For the organization, it means moving toward a more intelligent, connected, and easy-to-adopt automation experience.
Practical Example: A Purchase Request From an AI Assistant
Imagine a company has automated its purchase request process in Flokzu.
Before, a user had to enter the platform, find the process, complete the form, load the data, and submit the request. With the MCP connection, the user can ask their assistant:
“Create an instance of the Purchase Request process with an amount of 5,000 USD and ACME as the supplier.”
The assistant interprets the request, uses the corresponding tool from the MCP server, and creates the instance in Flokzu.
Then, the process follows its normal workflow: validations, approvals, business rules, alerts, responsible users, and traceability. The difference is how the action was started. The user was able to do it from a conversational interface without breaking process governance.
This example shows the real value of the integration: AI simplifies the interaction, while Flokzu maintains the structure, control, and workflow execution.
Key Takeaways
- Flokzu’s MCP server allows you to connect compatible AI assistants to your Flokzu account.
- You can interact with processes, databases, users, roles, and history using natural language.
- The integration makes it possible to start instances, check data, update records, manage users, and export information.
- The scope of available actions depends on the API Key and defined configuration.
- The connection uses HTTPS, does not persist API Keys, and maintains tenant isolation.
- MCP turns the AI assistant into a new interface for operating automated processes.
- Flokzu combines BPM, no-code, and AI to make complex process management simpler.
Frequently Asked Questions About Flokzu’s MCP Server
What is Flokzu’s MCP server?
It is a server compatible with Model Context Protocol that allows AI assistants to connect with Flokzu to execute actions, retrieve information, and operate on processes using natural language.
Which AI assistants can I connect?
You can use clients compatible with MCP over HTTP, such as Claude Desktop, Claude Code, Cursor, VS Code, and other agents that support MCP over HTTP.
What can I do from the AI assistant?
You can start process instances, check instance data, update fields, work with databases, manage users and roles, and export history or metadata from attached documents.
Do I need an API Key?
Yes. To configure the connection, you need an API Key and the email of the administrator user on whose behalf operations will be executed. The scope of the API Key defines what the assistant can do.
Is it secure to connect an AI assistant to Flokzu?
The connection to Flokzu’s MCP server uses encrypted HTTPS transport. The server does not persist API Keys, maintains tenant isolation, and records operations in Flokzu’s history associated with the configured email.
Does my data go to the AI provider?
According to Flokzu’s documentation, only the information the user writes and receives in the chat goes to the AI provider. Calls to the MCP server go directly from the local client to Flokzu, without passing through the AI provider’s servers.
Can I limit which tools the AI can use?
Yes. Most MCP clients allow you to enable or disable individual tools from their configuration. The API Key scope also helps limit which actions the assistant can execute.
What happens if my API Key changes or expires?
You must update the API Key value in the MCP client configuration file and restart the client. There is no persistent session.
Connect Your AI Assistant to Your Flokzu Processes
The way teams work with processes is changing. Today, organizations need automation, traceability, and control, but also simpler, faster, and more natural experiences.
With Flokzu’s MCP server, you can connect your AI assistant to your automated processes and start interacting with them using natural language. From starting a request to checking records, managing users, or exporting history, this integration opens a new way to operate on Flokzu.
AI becomes another interface for working with your processes. Flokzu maintains the control, rules, traceability, and execution.
Connect your AI assistant to Flokzu’s MCP server and discover a simpler way to interact with your automated processes.
