Step 2: Connect Google Search Console

Authorize Google Search Console access and prepare the HTTP request nodes that fetch properties and insights.

1. Open Fetch Search Console Insights and set URL to =https://www.googleapis.com/webmasters/v3/sites/{{ $json.property }}/searchAnalytics/query.
2. Set Method to POST, Specify Body to json, and JSON Body to ={ "startDate": "{{ $json.start_date }}", "endDate": "{{ $json.end_date }}", "dimensions": {{ $json.dimensions }}, "rowLimit": {{ $json.rowLimit }}, "startRow": 0, "dataState":"all" }.
3. Enable Send Headers and set Content-Type to application/json in Fetch Search Console Insights.
4. Credential Required: Connect your oAuth2Api credentials in Fetch Search Console Insights.
5. Open Retrieve Search Properties and set URL to =https://www.googleapis.com/webmasters/v3/sites.
6. Credential Required: Connect your oAuth2Api credentials in Retrieve Search Properties.

Step 3: Set Up Input Mapping and Tool Invocation

Normalize the incoming payload and build API request fields that drive the routing logic.

1. In Map Input Fields, add assignments for chatInput, sessionId, and date_message using ={{ $json.body?.chatInput || $json.chatInput }}, ={{ $json.body?.sessionId || $json.sessionId }}, and ={{ $now.format('yyyy-MM-dd') }}.
2. Configure Tool Invocation Trigger as the entry point for tool calls from the AI agent.
3. In Build API Request Fields, map fields using expressions: ={{ $json.query.parseJson().request_type }}, ={{ $json.query.parseJson().startDate }}, ={{ $json.query.parseJson().endDate }}, ={{ $json.query.parseJson().dimensions }}, ={{ $json.query.parseJson().rowLimit }}, ={{ $json.query.parseJson().startRow}}, and ={{ encodeURIComponent($json.query.parseJson().property) }}.
4. Connect Tool Invocation Trigger to Build API Request Fields, then to Route by Request Type.

Step 4: Configure the AI Conversation Layer

Set up the conversational agent with memory, the OpenAI model, and the Search Console tool integration.

1. Open Conversational Data Agent and set Text to =user_message : {{ $json.chatInput }} date_message : {{ $json.date_message }}.
2. Connect OpenAI Dialogue Model to Conversational Data Agent as the language model, and set Model to gpt-4o.
3. Credential Required: Connect your openAiApi credentials in OpenAI Dialogue Model.
4. Connect Postgres Memory Store to Conversational Data Agent as memory and set Table Name to insights_chat_histories.
5. Credential Required: Connect your postgres credentials in Postgres Memory Store.
6. Connect Search Console Tool Invoke to Conversational Data Agent as a tool and set Name to SearchConsoleRequestTool.
7. For AI tool nodes like Search Console Tool Invoke, add any required credentials on the parent workflow that it calls; the tool node itself does not store credentials.
8. Connect Map Input Fields to Conversational Data Agent, then to Return Webhook Reply.

Step 5: Configure Routing and Response Aggregation

Route requests by type and format the Search Console responses for the agent.

1. In Route by Request Type, create two outputs: custom_insights and website_list based on ={{ $json.request_type }} and ={{$json.request_type}}.
2. Connect the custom_insights output to Fetch Search Console Insights, then to Create Search Data Array, and finally to Aggregate Response Payload.
3. In Create Search Data Array, set searchConsoleData to ={{ $json.rows }}.
4. Connect the website_list output to Retrieve Search Properties, then to Create Site List Array, and finally to Aggregate Response Payload 2.
5. In Create Site List Array, set searchConsoleData to ={{ $json.siteEntry }}.
6. In both Aggregate Response Payload and Aggregate Response Payload 2, set Aggregate to aggregateAllItemData and Destination Field Name to response.

Step 6: Test and Activate Your Workflow

Run a manual test to validate the chat-to-insights flow and then activate the workflow for production use.

1. Click Execute Workflow and send a POST request to the Incoming Chat Webhook URL with a sample body containing chatInput and sessionId.
2. Confirm that Map Input Fields outputs chatInput and date_message, and that Conversational Data Agent produces a response.
3. Verify that tool calls through Search Console Tool Invoke trigger Tool Invocation Trigger and that Route by Request Type sends requests to the correct Search Console path.
4. Successful execution should end with Return Webhook Reply returning a conversational response enriched by Search Console data.
5. Toggle the workflow Active to enable production use.