AI/ML, Integration

Ava Voice AI Morning Brief Agent

Real-time LiveKit voice agent that reads a daily business brief, answers from a company knowledge base, pulls live data from a read-only internal API and weather service, and triggers a signed Claude webhook for deeper analysis, all by voice from a mobile web page.

Ava Voice AI Morning Brief Agent

Technologies Used

Tools and platforms used to build this solution.

LiveKit Agents
Python
Ubuntu
aiohttp

How It Works

A closer look at the workflow behind the results.

1

Secure Sign-In

The user opens the call page on a phone or laptop and enters the team passcode once. The server checks it and issues a short-lived LiveKit token for a private call room.

2

Voice Conversation

LiveKit sends Ava into the call. The user's speech is turned into text in real time, and a turn detector makes sure Ava doesn't cut them off mid-sentence.

3

Function Calling

The AI model decides which tool to use and fetches real data: the read-only company API, the knowledge base, and live weather and time. It never makes up numbers.

4

Claude Webhook

For deeper analysis, Ava sends a signed webhook to a Claude skill. It runs in the background while she keeps talking, and she reads the result when it arrives.

5

Spoken Answer and Hang-Up

Ava replies in a natural voice with a live transcript on screen. When the user is done she says goodbye, ends the call, and saves it to the call history.

6

Always-On Deployment

The agent and call page run as services on an Ubuntu server with HTTPS, so they restart by themselves after a crash or reboot.

Results & Impact

Measurable outcomes that demonstrate the value delivered.

11 function tools: company API, knowledge base, live weather, time, Claude analysis and hang-up
Daily brief combines 2 sources (company data and live weather) fetched in parallel in one tool call
Calls survive network drops: Ava waits up to 20 seconds for the caller to reconnect
10 automated conversation tests (simulations) check the agent's behavior before release

Business Impact

Managers no longer check several systems each morning: Ava gives them a short spoken brief on sales, delays, urgent tickets and meetings, hands-free, even while driving. They can ask follow-up questions in plain language, and longer Claude analyses run in the background so the conversation never stalls. Company data stays safe because Ava can only read, never change anything, and every Claude request is signed.

Interested in a similar project?

Let's discuss how we can build something great together.

Start Your Project