Welcome to the exciting world of Agentic Workflows on the .do platform! If you're looking to build intelligent, automated processes or offer your services as seamless software, you've come to the right place. This post will guide you through setting up your first Agentic Workflow and how to leverage the canonical reference.do documentation to make it happen.
Before we dive into the "how," let's briefly touch on the "what." .do empowers you to define and execute complex business logic as "Agentic Workflows." These workflows are composed of individual "Agents" that perform specific tasks. Think of it as building sophisticated processes where different intelligent components collaborate.
Furthermore, the .do platform allows you to expose these workflows and agents as "Services-as-Software" (SaaS). This means your business logic becomes accessible and consumable programmatically, opening up a wealth of integration and automation possibilities.
As you embark on your journey with .do, the reference.do documentation becomes your indispensable guide. It's the single source of truth for everything related to building, understanding, and leveraging Agentic Workflows and Services-as-Software on the platform.
Within the reference, you'll find:
Let's outline the basic steps for creating a simple Agentic Workflow and see how the reference.do documentation supports you at each stage.
Step 1: Define Your Workflow's Goal
What do you want your Agentic Workflow to achieve? For this example, let's imagine a simple workflow that takes a customer's name as input, uses an Agent to greet the customer, and then passes that greeting to another Agent for logging.
Step 2: Design Your Agents
Based on your workflow's goal, you'll need at least two Agents:
Step 3: Consult the reference.do for Agent Definition
Now, head over to reference.do and look for the documentation on Agent Definition or Agent Models. You'll find details on how to structure your Agent’s definition, including inputs, outputs, and execution logic. This is where you'll learn about the expected structure for your Agent's code or configuration.
Step 4: Consult the reference.do for Workflow Definition
Next, explore the documentation related to Agentic Workflow Definition. This section in reference.do will show you how to sequence your defined Agents, pass data between them, and handle conditional logic. You'll learn about the workflow input/output structure and how to link the outputs of one Agent to the inputs of another.
Step 5: Implement Your Agent Logic
Depending on how you're building your Agents (.do supports various methods), you'll write the actual code or configure the behavior of your GreetingAgent and LoggingAgent. Refer back to the Agent API and Data Model sections in reference.do if you need clarity on how your Agent should receive input and structure its output.
Consider the structure of Agent step output as defined in the reference:
interface AgentStepOutput {
status: 'success' | 'failure';
output: any; // The output of the agent step can be anything
agentId: string; // The ID of the agent that executed this step
}
Understanding this structure is crucial for your Agents to communicate effectively within the workflow.
Step 6: Deploy Your Workflow
Follow the deployment instructions provided in the reference.do documentation. This will likely involve using the .do platform's tools or APIs to register your Agents and deploy your defined workflow.
Step 7: Execute Your Workflow (and check the reference!)
Once deployed, you can trigger your workflow. The API Documentation in reference.do will provide the necessary endpoint information and request format to initiate your workflow execution.
Review the documentation on Workflow Execution Output. The reference provides a standard structure for the overall workflow output:
interface AgentWorkflowOutput {
status: 'completed' | 'failed';
steps: AgentStepOutput[];
finalOutput: any; // The final output of the entire workflow
}
This helps you understand the result of your workflow execution, including the status and the output of each step.
We've compiled some common questions about the reference documentation:
Q: What is the purpose of the reference documentation?
A: The reference provides canonical documentation for all aspects of the .do platform, including API specifications, data models, and best practices for building Agentic Workflows and Services-as-Software.
Q: What kind of information is included in the API documentation?
A: You can find detailed information on API endpoints, request/response formats, data schemas, and code examples for interacting with the .do platform.
Q: Does the reference cover data models used on the platform?
A: Yes, the reference includes documentation on the underlying data models used within the .do platform, such as the structure of Agents, Workflows, and Services.
Q: Are there examples of building Agentic Workflows and Services-as-Software?
A: You can find guides and examples demonstrating how to implement common Agentic Workflows and build Services-as-Software using the provided APIs and SDKs.
Building your first Agentic Workflow on .do is an exciting step towards automating your business processes and offering services programmatically. The reference.do canonical reference is your essential companion throughout this journey, providing the technical details and guidance you need to succeed. Dive into the documentation, explore the API specifications, and start building your intelligent workflows today!
Ready to get started? Explore the comprehensive reference documentation at reference.do!