The Split by AI action uses LLMs to categorize incoming messages and route contacts down different paths in your flow.

Instead of writing rigid keyword rules, you define categories in plain language and let the AI handle typos, slang, and context.

Use this action when you need to make a routing decision based on the meaning of a contact's response.

Configuration

1. Split by AI: Select the Split by AI action, then choose the Model that will power it.
2. Input: Enter the text to classify (usually @input for the contact's last reply).
3. Categories: Enter the categories you want the AI to choose from, such as Course Question and Account Issue. You can give each one a description to help the model decide when it fits.
4. Require confidence (Optional): Choose how confident the AI must be, such as medium. Anything it classifies with lower confidence goes to Other.
5. Save the result as (Optional): Enter a name to save the chosen category to the flow context (e.g., Question Type).

In the example below, a student asks "Do I need BIO 101 first?" Split by AI sorts the question as a Course Question, and a Call AI action answers it from the advising FAQ.

Split by AI vs. Call AI

Split by AI is for routing. It makes the AI pick one category from a fixed list and builds a branch in your flow for each category. * Use this to decide where a contact goes next.

Call AI is for generation. You write a custom prompt, and the AI generates text, answers questions, or summarizes a conversation. It does not create branches. * Use this when you want AI to write something.

Use Cases

• Support Triage: Classifying messages into Billing, Tech Support, or Sales to route them to the right team.
• Sentiment Analysis: Sorting feedback into Positive, Neutral, or Negative to trigger the right follow-up.
• Survey Classification: Grouping open-ended answers into themes like Price, Features, or Usability.