Skip to main content

Analyze sentiment with AI

Analyze sentiment with AI reads a block of text and classifies its tone — positive, neutral, or negative. Each sentiment creates a separate branch in your automation, so different tones trigger different follow-up actions. The AI also returns a short reason explaining its classification. No code, no keyword lists, no manual triage.

How it works​

  1. You provide input text — the content the AI should analyze (a message, review, form response, etc.)
  2. The AI reads the input and classifies the sentiment as positive, neutral, or negative
  3. The automation continues down the matching sentiment's branch
  4. A Reason output explains why the AI chose that sentiment

If the AI cannot determine the sentiment, it returns an empty result and the automation follows the default path.

Set up the action​

  1. Open the automation builder and add a new step
  2. Under AI actions, select Analyze sentiment with AI
  3. Configure the fields:

1. Provide the text to analyze​

In the Input text field, enter or insert the content you want analyzed. Select Add dynamic content to insert data from a previous step — for example, a customer message from a trigger, a review, or a form response.

tip

Dynamic content values are resolved before the AI reads the text, so the AI always sees the actual data — not the placeholder.

2. Customize the categories (optional)​

Expand View or edit AI configuration to see or modify the sentiment categories. The defaults handle most use cases:

CategoryDescription
positiveThe text expresses satisfaction, enthusiasm, gratitude, or approval
neutralThe text is factual or informational, or the tone is unclear or mixed
negativeThe text expresses dissatisfaction, frustration, anger, or disapproval

You can add, remove, rename, or reorder categories to match your workflow. Each category creates a separate branch in the automation.

info

Each category you define creates a separate branch. After classification, the automation continues down the matching branch only.

Outputs​

After the step runs, two values are available to downstream steps:

OutputDescription
SentimentThe matched category key (e.g., "positive", "neutral", "negative"). Empty if the text could not be classified.
ReasonA one-sentence explanation of why the AI chose that sentiment.

Reference these outputs in later steps using Add dynamic content and selecting this step's outputs.

Tips for effective sentiment analysis​

  • Use the defaults first — The built-in positive/neutral/negative categories work well for most customer-facing text. Only customize if you have a specific need.
  • Provide enough text — A single word like "fine" is ambiguous. The more context the AI has, the more accurate the classification.
  • Add custom categories sparingly — Too many categories make the AI's job harder. 3 to 5 well-defined categories produce the most consistent results.
  • Use the Reason output — Include it in notifications or CRM notes so the person acting on the sentiment understands why the AI flagged it.

Use case examples​

Escalate negative AI employee conversations​

Route frustrated customers to a senior rep while positive and neutral conversations follow the standard path.

Step 1 — Set the trigger

Select Webchat captures a lead.

Step 2 — Add the Analyze sentiment with AI step

  • Under AI actions, select Analyze sentiment with AI
  • Input text: select Add dynamic content and choose Chat Summary from the trigger

Step 3 — Configure each branch

  • Negative — Add a Notify a salesperson step targeted at a team lead, with the Reason output in the notification body and a link to the full conversation
  • Positive / Neutral — Add a Notify a salesperson step with standard follow-up routing to the contact's owner
tip

Include the Reason output in the notification body so the rep knows why the conversation was flagged — e.g., "Customer expressed frustration about repeated billing errors and threatened to cancel."

Route form submissions by customer tone​

Send angry customers to immediate human attention, log neutral inquiries, and funnel happy customers into referral campaigns.

Step 1 — Set the trigger

Select A form is submitted for a contact (under Contacts).

Step 2 — Add the Analyze sentiment with AI step

  • Under AI actions, select Analyze sentiment with AI
  • Input text: select Add dynamic content and choose the form field containing the customer's message

Step 3 — Configure each branch

  • Negative — Add a Create a CRM sales task for the contact step marked urgent, assigned to the account manager, with the Reason in the description
  • Neutral — Add a Create a CRM note step to log the submission on the contact record
  • Positive — Add a Start a campaign for the contact step that enrolls them in a referral or testimonial request campaign

Prioritize AI employee handoffs by customer mood​

Automatically set task priority based on whether the customer was frustrated or calm during the AI employee conversation.

Step 1 — Set the trigger

Select Webchat captures a lead or Call activity added with a conditional filter for AI voice receptionist.

Step 2 — Analyze sentiment

  • Under AI actions, select Analyze sentiment with AI
  • Input text: select Add dynamic content and choose the chat summary

Step 3 — Create prioritized tasks

  • Negative — Add a Create a CRM sales task for the contact step with high priority, assigned to a senior rep, description: "Negative sentiment — [Reason]"
  • Neutral / Positive — Add a Create a CRM sales task for the contact step with normal priority, assigned to the contact's owner
tip

Agents see at a glance which contacts need careful, urgent handling and which are routine follow-ups — no more guessing which tasks to pick up first.