AI decision

Type: llm.

#What it does

1. Builds the instructions the same way as AI reply, plus emotion cues if emotion is on.
2. Asks the AI for a JSON answer with the fields in llm_return_type, and waits for the whole answer.
3. Saves every returned field to the call's state (a plain-text answer is saved as llm_response), and the exchange to memory.
4. Says nothing. Follow it with a Say that speaks speak, and a Branch if the call should go different ways.

It waits for the full answer before anything is said, so it's slower to the first word than AI reply. Use it where you need the fields; use AI reply for plain chat.

#Settings

Everything goes in the step's parameters. In the editor, click the step to open its settings.

system_promptstring

In the editor: Instructions. Tell the agent who it is, what to do and how to talk. Default: "".

modelstring

In the editor: Model. Which AI writes the reply. The provider comes with it. Default: "openai/gpt-oss-120b".

llm_return_type{ name: { type, description } }

In the editor: What it decides (More tab). Things the AI fills in every turn. The spoken reply is what the caller hears; the rest goes to your webhook.

emotionboolean

In the editor: Add emotion. Lets the voice sound happy, calm or excited where it fits.

emotion_ttsstring

In the editor: Emotion style. Which voice understands the emotion cues. Only used when emotion is on. One of: eleven_v3 (ElevenLabs v3).

input_variables{ name: { type, description } }

In the editor: What it reads (More tab). Information from earlier steps the AI can use.

history_keystring

In the editor: Shared memory (More tab). Steps with the same name here remember the same conversation. Default: "conversation_history".

summariserboolean

In the editor: Summarise long calls (More tab). Keeps long conversations quick by summarising older turns.

prompt_templatestring

Built-in instructions to start from. See Prompt templates. Default "base_llm".

summariser_configobject

Tune summariser. See Memory.

#Examples

1{
2  "type": "llm",
3  "parameters": {
4    "system_prompt": "Find out if the caller wants a demo this week.",
5    "service": "groq",
6    "model": "openai/gpt-oss-120b",
7    "llm_return_type": {
8      "speak": {
9        "type": "str",
10        "description": "What to say."
11      },
12      "hangup": {
13        "type": "bool",
14        "description": "True once done."
15      },
16      "outcome": {
17        "type": "str",
18        "description": "demo_booked, later or not_interested"
19      }
20    },
21    "history_key": "conversation_history"
22  },
23  "next": "reply"
24}

#Next steps

• Steps: all the other steps.
• Patterns: see it in a full workflow.