Your First Workflow
This tutorial walks you through building a complete, working workflow from scratch. By the end, you will have a pipeline that fetches patients from your Epic system, sends their clinical data through an AI summarizer, and produces a structured PDF report.
What you will build
A 3-node sequential pipeline:
- Epic Patient List — Fetches all active patients from your Epic connection
- Clinical Summarizer — Uses an AI model to generate a brief clinical summary for each patient
- Report Generator — Compiles the summaries into a downloadable PDF report
This is one of the most common starting workflows because it demonstrates data ingestion, AI processing, and structured output — the three fundamental patterns in TietAI.
Prerequisites
Before starting, confirm:
- You have logged in and completed your profile (Quick Start)
- Your organization has an Epic connector configured — the connector name appears in Settings → Connectors with a green status indicator (Connect Your EHR)
- Your role is Clinician or Admin (Viewers cannot create or run workflows)
If you do not have an Epic connector, you can still follow this tutorial using a Generic FHIR R4 connector pointed at a test FHIR server, or contact your administrator to set up the Epic connection first.
Step 1: Open the Hydra Builder
From the main navigation bar on the left, click Hydra Builder. The studio opens with a blank canvas in the center, the node palette on the left, and the toolbar across the top.
If you see existing workflows listed, you are on the Workflows overview page — click New Pipeline in the top-right corner to open the canvas.
Step 2: Name your workflow
A dialog appears asking for a workflow name and optional description.
- Name:
Daily Patient Summary - Description:
Fetches active patients from Epic and generates a clinical summary report.
Click Create. The canvas opens with an empty workspace.
Use names that describe what the workflow does and how often it runs. "Daily Patient Summary" is clear; "Test Pipeline 3" is not. You will thank yourself when you have 20 workflows.
Step 3: Add the Epic Patient List node
In the node palette on the left, locate the Healthcare category. You can either scroll to it or type epic in the search bar at the top of the palette.
Find Epic Patient List and drag it onto the canvas. Drop it roughly in the left-center of the canvas — this will be the first node in your pipeline.
Configure the node:
Double-click the node to open its configuration dialog.
| Field | Value | Notes |
|---|---|---|
| Connector | Select your Epic connector from the dropdown | The dropdown shows all connectors in your org |
| Operation | search | Fetches a list of patients |
| Resource type | Patient | Pre-filled |
| Search parameters | Leave empty | Returns all active patients |
| Batch size | 10 | Patients per batch; start small for testing |
Click Save Configuration. The node header updates to show the connector name.
What to expect: This node will produce a fhir-bundle output containing Patient resources — one per matching patient in your Epic system.
Step 4: Add the Clinical Summarizer node
In the node palette, locate the AI / ML category. Find Clinical Summarizer and drag it onto the canvas, placing it to the right of the Epic Patient List node with some space between them.
Connect the nodes:
Hover over the Epic Patient List node. A small circle appears on its right edge — this is the output port. Click and drag from this port to the input port on the left edge of the Clinical Summarizer node. A line connects them.
The line color indicates data type compatibility. A solid green line means the port types match (fhir-bundle → fhir-bundle). If the line is red, the port types are incompatible and the connection will be rejected.
Configure the Clinical Summarizer:
Double-click the Clinical Summarizer node.
| Field | Value |
|---|---|
| Prompt | For each patient in this bundle, write a 2-sentence clinical summary covering their primary active conditions and current medications. Return a list in the same order as the input. |
| Model | Leave as default (your org's configured LLM) |
| Temperature | 0.2 (lower = more consistent, factual output) |
| Max tokens per patient | 200 |
Click Save Configuration.
Step 5: Add the Report Generator node
In the node palette, find Report Generator under the Output category. Drag it to the right of the Clinical Summarizer node.
Connect the Clinical Summarizer's output port to the Report Generator's input port.
Configure the Report Generator:
Double-click the node.
| Field | Value |
|---|---|
| Report title | Daily Patient Summary — {{date}} |
| Output format | PDF |
| Template | Standard Clinical Report |
| Include metadata | Checked |
The {{date}} token is replaced automatically with the execution date when the report is generated.
Click Save Configuration.
Step 6: Save the workflow
Click the Save button in the toolbar (or press Ctrl+S). The toolbar shows a "Saved" confirmation. TietAI saves automatically as you work, but it is good practice to save explicitly before running.
Your canvas should now show three nodes connected left-to-right: Epic Patient List → Clinical Summarizer → Report Generator.
Step 7: Run the workflow
Click the Run button in the top toolbar. A confirmation dialog appears asking if you want to run immediately or schedule a future run — select Run now.
The execution panel opens at the bottom of the screen. You will see:
- Epic Patient List — status changes from Pending to Running. A counter shows patients fetched.
- Clinical Summarizer — activates once the patient data arrives. This node may take 30–60 seconds depending on how many patients are in your Epic system.
- Report Generator — runs last, compiling the summaries.
Each node shows a green checkmark when it completes successfully.
Verifying it worked
When all three nodes show green:
- The execution panel shows status Completed with a duration timestamp.
- Click the Report Generator node in the execution panel to see its output — you will see a link to the generated PDF.
- Navigate to Reports in the main navigation. Your new report appears at the top of the list, titled with today's date.
- Click the report to preview it in-browser, or click Download PDF.
What to do if it fails
Epic Patient List fails with "Connector error"
The Epic connection timed out or credentials expired. Go to Settings → Connectors, find your Epic connector, and click Test Connection. If it fails, re-enter your credentials. See Connect Your EHR for details.
Clinical Summarizer fails with "Model unavailable"
Your organization's LLM endpoint may be temporarily unavailable. Wait 2 minutes and click Retry on the failed execution. If it persists, contact your TietAI administrator.
Clinical Summarizer fails with "Input type mismatch"
The connection between Epic Patient List and Clinical Summarizer is using an incompatible port type. Delete the connection line, then re-draw it. Ensure the line is green before saving.
Report Generator fails with "Template not found"
The "Standard Clinical Report" template may not be installed in your organization. Select a different template from the dropdown, or contact your administrator to install it.
The execution panel shows no activity
If the execution status stays "Pending" for more than 30 seconds, the workflow engine may be starting up. Refresh the page — if the execution is not listed at all, try running again.
Next steps
You have built and run your first TietAI workflow. Here is where to go next:
- Add scheduling — Click the dropdown arrow next to Run and select Schedule. Set it to run every day at 7:00 AM so the report is ready at the start of each shift.
- Explore more nodes — Browse the Hydra Builder reference to see all available node categories and what each one does.
- Connect more systems — Add a Cerner connector or a FHIR write node to route summarized data back to your EHR.
- Set up notifications — Add a Notification node after the Report Generator to send an email or in-app alert when the report is ready.