Step 2: Capture and Route the Incoming Message

These nodes store the user message and route it into the research flow.

1. In Capture Chat Input, set the assignment User Chat Message to {{ $json.chatInput }}.
2. Open Route Valid Prompt and keep both rules that check non-empty values: {{ $json['Retrigger Chat'] }} and {{ $json['User Chat Message'] }}.
3. Ensure Route Valid Prompt outputs into Research Orchestrator (both output routes connect to the same node).
4. In Repeat Prompt Builder, keep Retrigger Chat set to the full prompt string that requests the next batch of 10 prospects.

Step 3: Set Up the AI Research Orchestrator

The agent coordinates the chat model, research tools, and data writes.

1. Open Research Orchestrator and set Text to {{ $json['User Chat Message'] }} {{ $json['Retrigger Chat'] }}.
2. Verify Research Orchestrator uses the built-in system message that defines Dream 100 criteria (already configured).
3. Open OpenRouter Dialogue Model and connect it as the language model for Research Orchestrator via the existing AI language model connection.
   Credential Required: Connect your openRouterApi credentials.
4. Open Session Memory Buffer to keep conversation context; no credentials needed. This memory connects to Research Orchestrator and should remain linked.
5. Open Perplexity Research Tool and keep Model set to sonar-deep-research.
   Credential Required: Connect your perplexityApi credentials.
   Note: This tool is used by Research Orchestrator, so credentials are added here but consumed by the agent.
6. Open Upsert Sheet Records and keep Operation set to appendOrUpdate with Matching Columns set to Prospect ID.
   Credential Required: Connect your googleSheetsOAuth2Api credentials.
   Note: This tool is invoked by Research Orchestrator, so ensure it stays connected as an AI tool.

Step 4: Connect Google Sheets for Prospect Tracking

The workflow checks the sheet for completion and appends or updates rows.

1. In Fetch First Row, set Document to your sheet ID (replace [YOUR_ID]) and Sheet to Sheet1 (gid=0).
2. Keep the filter in Fetch First Row set to Prospect ID = 1.
   Credential Required: Connect your googleSheetsOAuth2Api credentials.
3. In Fetch Last Row, set the same Document and Sheet values and keep the filter Prospect ID = 100.
   Credential Required: Connect your googleSheetsOAuth2Api credentials.
4. In Store First Name, set Prospect 1 to {{ $json['Prospect Name'] }}.
5. In Store Last Name, set Prospect 100 to {{ $json['Prospect Name'] }}.
6. Verify the execution flow: Research Orchestrator → Fetch First Row → Store First Name → Fetch Last Row → Store Last Name → Check Sheet Empty.

Step 5: Configure Conditional Logic and Chat Responses

These nodes handle empty or partially completed sheets and continue research cycles.

1. In Check Sheet Empty, keep both conditions checking for empty values: {{ $('Store First Name').item.json['Prospect 1'] }} and {{ $('Store Last Name').item.json['Prospect 100'] }}.
2. In Request More Details, set Message to {{ $('Research Orchestrator').item.json.output }} and keep Wait User Reply set to false.
3. In Check Sheet Completion, keep the logic that checks Prospect 1 exists and Prospect 100 is empty.
4. Ensure Check Sheet Completion routes to Repeat Prompt Builder (for continuing research) and to Send Completion Note when the sheet is complete.
5. In Send Completion Note, update the Message field from = to a real completion message before production use.

Step 6: Configure Slack Utility Nodes (Optional)

These utility nodes can be used to request more details or notify completion in Slack.

1. Open Utility: Slack Message Trigger and select the Channel ID you want to listen to.
2. Open Utility: Slack Request Details and configure the message payload you want to send.
3. Open Utility: Slack Completion Notice and configure the completion message content.

Step 7: Test and Activate Your Workflow

Validate the end-to-end flow before enabling it in production.

1. Click Execute Workflow and send a sample chat message through Chat Intake Trigger.
2. Confirm that Capture Chat Input populates User Chat Message and that Route Valid Prompt sends data to Research Orchestrator.
3. Verify Fetch First Row and Fetch Last Row read from your sheet, and that Upsert Sheet Records writes or updates rows.
4. Check that Request More Details or Send Completion Note posts a chat response based on sheet status.
5. When everything works, toggle the workflow to Active to run it in production.