The AI That Calls You: Setting Up a Proactive Voice Agent

How to Set Up a Proactive AI Voice Agent

The other day I was watching a YouTube video, learning about AI agents, and this one really resonated with me—an agent you interact with via a phone call. Instead of constantly checking dashboards or messaging a chatbot, this system actually calls your phone to give you morning briefings, business metric updates, or even simple daily reminders.

Why Use a Voice Agent?

Text notifications are easily ignored, but phone calls demand immediate attention. An AI voice agent acts proactively. It can serve as a 24/7 receptionist for a business, a personal chief of staff, or an automated accountability partner that checks in on your daily goals.

The Setup Tutorial

This system relies on two main components: OpenClaw (the brain that fetches data and makes decisions) and Synthflow AI (the voice platform that handles the actual phone call).

  1. Configure Synthflow AI: Create an account and build a new "Outbound Voice Agent". Select the Prompt Builder, give your agent a name, and choose your AI model (like GPT-4o) alongside your timezone.
  2. Customize the Voice: Pick an accent and adjust the talking speed and volume to fit your preference.
  3. Add Knowledge: Upload PDFs containing your specific business or personal information to the agent's Knowledge Base.
  4. Enable Memory: Turn on transcript and recording saving so the AI remembers past conversations. Add a "Memory Group" so it recalls previous calls and builds context with you over time.
  5. Connect to OpenClaw: Purchase a phone number within Synthflow. Grab your API Key and Agent ID, then feed them into your OpenClaw prompt to link the logical brain with the voice synthesizer.

Security & Cost Breakdown

  • Security: Always set caller ID filtering to reject unknown numbers. Set a maximum call duration (e.g., 300 seconds) so the AI doesn't run up a bill if something glitches.
  • Cost: Running this exact setup costs roughly $35–$45 per month ($29/mo for Synthflow, plus $5–$15/mo for the LLM API usage).

Resources Used:

• Older →