How to Monitor Slack Feedback with OpenAI Sentiment Analysis using N8n

Workflow > Add Node > Slack Trigger

Set up Slack trigger node

Create a new workflow and configure the Slack trigger to monitor your feedback channel. This node will fire whenever a new message is posted to the specified channel.

1. 1Click the + button to add a new node
2. 2Search for 'Slack' and select 'Slack Trigger'
3. 3Choose 'On New Message Posted to Channel' from the events dropdown
4. 4Connect your Slack workspace using OAuth2

Slack Trigger > Parameters

Configure channel monitoring

Specify which channel to monitor and set up message filtering. This prevents the workflow from analyzing bot messages or system notifications.

1. 1In the Channel field, type '#feedback' or select from the dropdown
2. 2Toggle 'Listen to Bot Messages' to OFF
3. 3Set 'Include Threads' to YES if you want to analyze thread replies
4. 4Click 'Test Step' to verify the connection

Workflow > IF Node > Conditions

Add message filtering logic

Create an IF node to filter out unwanted messages like empty posts, URLs-only messages, or messages shorter than 10 characters. This saves API calls to OpenAI.

1. 1Add an IF node after the Slack trigger
2. 2Set condition to 'String' > 'Length' > 'Larger' > '10'
3. 3Map the input to {{$node['Slack Trigger'].json['text']}}
4. 4Connect the 'true' output to continue the workflow

Workflow > Add Node > OpenAI > Chat

Connect OpenAI node

Add the OpenAI node and configure it for text analysis. You'll use the chat completion endpoint with GPT-3.5-turbo for fast, cost-effective sentiment analysis.

1. 1Add an OpenAI node after the IF node's true output
2. 2Select 'Chat' as the resource type
3. 3Choose 'Create a Chat Completion' as the operation
4. 4Add your OpenAI API key in the credentials section

OpenAI Node > Parameters > Messages

Configure sentiment analysis prompt

Set up the system prompt to ensure consistent sentiment classification. The prompt instructs GPT to return only 'positive', 'negative', or 'neutral' for reliable parsing.

1. 1Set Model to 'gpt-3.5-turbo'
2. 2In System Message, enter: 'Analyze the sentiment of this customer feedback. Respond with only one word: positive, negative, or neutral.'
3. 3Map User Message to {{$node['Slack Trigger'].json['text']}}
4. 4Set Max Tokens to 10 and Temperature to 0

Workflow > Add Node > Code > JavaScript

Add sentiment parsing logic

Use a Code node to extract and clean the sentiment result from OpenAI's response. This handles edge cases where GPT returns extra words or formatting.

1. 1Add a Code node after the OpenAI node
2. 2Select JavaScript as the language
3. 3Add code to parse the sentiment from the OpenAI response
4. 4Include logic to default to 'neutral' if parsing fails

Workflow > Add Node > Switch > Routes

Set up Slack response routing

Create a Switch node to route different sentiment types to different Slack actions. This lets you customize the response based on whether feedback is positive, negative, or neutral.

1. 1Add a Switch node after the Code node
2. 2Create three routes: positive, negative, neutral
3. 3Set the input value to {{$node['Code'].json['sentiment']}}
4. 4Configure each route to match the exact string values

Switch > Positive Route > Slack Node

Configure positive feedback actions

Set up what happens when positive feedback is detected. This typically involves posting to a different channel or adding reaction emojis to celebrate good feedback.

1. 1Connect a Slack node to the 'positive' output
2. 2Choose 'Add Reaction' as the operation
3. 3Map the channel and timestamp from the trigger
4. 4Set the reaction emoji to 'thumbsup' or 'tada'

Switch > Negative Route > Slack Node

Handle negative feedback alerts

Configure escalation for negative feedback by posting to a management channel or mentioning specific team members. Critical feedback needs immediate attention.

1. 1Connect another Slack node to the 'negative' output
2. 2Choose 'Post Message' as the operation
3. 3Set channel to '#customer-success' or your escalation channel
4. 4Create a message template with the original feedback and user

Switch > Neutral Route > Slack Node

Log neutral feedback

Set up simple logging for neutral feedback that doesn't require immediate action but should be tracked for analysis. A simple emoji reaction works well.

1. 1Connect a Slack node to the 'neutral' output
2. 2Choose 'Add Reaction' as the operation
3. 3Map channel and timestamp fields
4. 4Set reaction to 'eyes' or 'memo' to indicate processing

Workflow > Execute Workflow

Test the complete workflow

Run end-to-end tests with sample messages to verify sentiment classification and routing work correctly. Test all three sentiment types to confirm each path executes properly.

1. 1Click 'Execute Workflow' in the top right
2. 2Post a test message to your #feedback channel
3. 3Check that the workflow triggered and completed successfully
4. 4Verify the correct Slack action was taken based on sentiment

Workflow > Settings > Error Handling

Activate continuous monitoring

Enable the workflow to run automatically on new messages. Set up error handling to prevent workflow failures from breaking ongoing monitoring.

1. 1Click the 'Active' toggle in the top right
2. 2Go to Settings > Error Handling
3. 3Set 'Continue on Fail' to ON for all nodes
4. 4Configure webhook timeout to 30 seconds

Drop this into an n8n Code node.

// Clean and validate GPT sentiment response
const response = $node['OpenAI'].json['choices'][0]['message']['content'].toLowerCase().trim();
const sentiment = ['positive', 'negative', 'neutral'].find(s => response.includes(s)) || 'neutral';

// Clean and validate GPT sentiment response
const response = $node['OpenAI'].json['choices'][0]['message']['content'].toLowerCase().trim();
const sentiment = ['positive', 'negative', 'neutral'].find(s => response.includes(s)) || 'neutral';
return [{json: {sentiment, confidence: response.length < 20 ? 'high' : 'low'}}];