To have an agent speak immediately when a call starts, without waiting for the caller to speak, you can use one of the following methods: • Using `session.say()`: This method allows the agent to speak a predefined message immediately. For example:

await agent.say("Hi there, how are you doing today?", allow_interruptions=True)

• Using `session.generate_reply()`: This method enables the agent to generate and speak a response based on specific instructions. For example:

await session.generate_reply(instructions="say hello to the user")

Implementation Examples: 1. *In the Agent Class (

on_enter

method):*`class MyAgent(Agent): async def on_enter(self): self.session.generate_reply(instructions="say hello to the user")` 2. *After the Agent Starts:*`agent.start(ctx.room) await agent.say("Hi there, how are you doing today?", allow_interruptions=True)` 3. For Telephony (Inbound Calls): If it’s an inbound call, you can ensure the agent speaks first:

if phone_number is None:
    await session.generate_reply(
        instructions="Greet the user and offer your assistance."
    )

Key Notes: • By default, the agent waits for user input before responding. • To initiate a conversation, use the above methods to greet users at session start. • If using

session.say()

, ensure a TTS plugin is added to your session. Alternatively, use

session.generate_reply()

for a more flexible approach. These methods will allow the agent to proactively greet users as soon as the call begins. Sources: Agent speech and audio | LiveKit Docs | examples/voice_agents/preemptive_generation.py | Agents telephony integration | LiveKit Docs