Build a Wingz Workflow Using the Wingz Workflow Assistant

Overview

Building a Wingz workflow normally means placing a Start and End node, connecting Agent nodes, and wiring in any logic nodes on the canvas. The Wingz Workflow Assistant offers a faster path to the same result: describe the automation you want in plain language, and the assistant assembles the workflow for you, choosing and configuring the trigger, agent nodes, and any logic nodes the request calls for.

The assistant doesn't just place and connect nodes on the canvas. After each request, it explains what it built and why, so you can review its choices before publishing rather than having to infer them from the canvas alone. You can also continue the conversation to refine what it just built, the same way you'd keep editing a workflow manually.

This article covers building a new workflow with the Wingz Workflow Assistant, refining a workflow through follow-up messages, scoping a change to a specific node, and reviewing the assistant's edit history.

For more information on the manual approach to building a workflow, see Build a Wingz Workflow Manually to Automate Tasks. For an overview of the nodes the assistant configures on your behalf, see Introduction to Wingz Workflows.

Launch the Wingz Workflow Assistant

The Wingz Workflow Assistant is built into the Workflow canvas as a chat panel on the left, so it's available whether you're starting a workflow from scratch or editing one you've already built.

To launch the Wingz Workflow Assistant:

1. Log in to your Wingify account.
2. From the left navigation panel, go to Wingz > Workflows. This opens the Workflows dashboard.
3. Open the Workflow canvas by doing one of the following:
   1. Click Create to start from a blank canvas.
   2. Click an existing workflow in the list to continue working on it.

The assistant panel opens automatically with the Workflow editor. For a new workflow, it greets you with a few example prompts you can click to get started, such as summarizing running campaigns or reporting on paused campaigns. For an existing workflow, the same assistant panel opens alongside it. You can continue refining the workflow through further messages within the panel.

Build a Workflow from a Prompt

Once the assistant panel is open, building a workflow involves describing what you want, reviewing what the assistant sets up, and optionally rating that response before you refine it further.

Step 1: Describe the Workflow/Automation

1. In the message input box at the bottom of the assistant panel, describe the workflow in a single message. A useful description covers what should trigger it, what it should analyze or do, and who should be notified. For example:
   1. Create a workflow that summarizes all my running campaigns every month, and emails me a PDF report.
   2. Summarize paused campaigns and recommend whether to relaunch, iterate, or archive them.
      
      Alternatively, click one of the ready-to-use sample prompts to start instead of writing your own. Or, use the voice recorder icon in the message box to describe the workflow by speaking instead of typing. You can also attach files, reference specific agents, or bring in connectors as part of your message, so the assistant has more context to work with beyond the text you type.

2. Press Enter or click the send icon . The assistant responds with a short confirmation, followed by an expandable Reasoning section and a plain-language summary of what it set up, such as the trigger schedule, the Agent's task, and the notification recipient.

   Note: While the assistant is reasoning through your request, you can click the stop icon to cancel the response if you no longer need it, or click Notify to get notified once the response is ready.

Step 2: Review the Generated Workflow

Once the assistant finishes responding, the draft workflow is created on the canvas on the right. Expand the Reasoning section in the assistant chat panel to see how it interpreted your request. 

The chat panel displays a workflow card with the workflow's name, a Workflow Editor label, and the Active status tag.

Step 3: Give Feedback (Optional)

Use the thumbs up , thumbs down , or copy icons below any response to rate or reuse its output. This feedback does not change the workflow. To refine the workflow, send another message.

Refine a Workflow Through Follow-Up Messages

Building a workflow with the assistant is rarely a single message. Once you've reviewed what it built, continue the same conversation to adjust it. Each follow-up message is treated as an edit to the current version, not a request to start over, so the assistant builds on what's already on the canvas instead of replacing it.

Step 1: Send a Follow-Up Instruction

Describe the change you want, the same way you described the workflow automation. The assistant can, for example:

• Add a Condition node to control when the notification is sent based on a value in the Agent's output.
• Add Duplicate and Merge nodes to run two Agent tasks in parallel and combine their results.
• Change the notification recipient, schedule, or an Agent's instructions.

The assistant responds the same way it does to a new request: a short confirmation, a Reasoning section, and an updated summary reflecting the workflow's logic after the change, not just the part you asked it to edit.

Step 2: Modify a Specific Node

A follow-up instruction can be ambiguous once a workflow has several similar nodes, for example, two Condition nodes or three parallel Agent branches.

To modify a specific node:

1. Click the node selector icon in the message box.
2. Click the node on the workflow canvas you want to reference. The node appears as a chip above your message, for example condition_1. 
3. Add your instruction and hit Enter or click Send.

The assistant applies the change only to the selected node.

Step 3: Review and Save the Update

After each edit, a banner with the message Workflow updated successfully. Review the changes before saving. appears at the top of the canvas. 

Review the updated nodes on the canvas to confirm the change matches what you asked for. If it doesn't, send another message describing the correction rather than adjusting the nodes manually, so the conversation maintains an accurate record of what was asked for and applied.

Review and Apply Past Assistant Edits

A workflow built through several rounds of assistant edits can end up structured quite differently from how it started, since each follow-up message changes the canvas rather than starting fresh. If you are unsure which message caused a change, or want to verify that an instruction was applied correctly, check the edit history. Every edit the assistant makes is recorded individually, and you can open any of them to see exactly what that edit changed.

To review and apply a past edit:

1. Scroll through the conversation in the assistant panel to the workflow version card for the edit you want to check. Each card shows the workflow version’s name, a Workflow Editor label, and a concise summary of the changes made in that version.
2. Click View Version to open that version in the editor. The card's status tag updates to Active, indicating it is the version currently loaded in the editor.

3. Review what changed in that specific edit, for example a node that was added, removed, or reconfigured.

   Tip: Use version history to catch unintended changes early. If an edit isn't what you intended, correct it with a follow-up message before making further changes.
4. If the version you're viewing is the one you want to keep, click Save to make it the workflow's current saved state, then continue with your normal Test and Publish steps to take it live.

Need more help?

For more information or further assistance, contact Wingify Support.