Lost & Found Caller — CALL-E

An AI voice agent that calls lost-and-found departments on your behalf, describes your missing item, checks for matches, and helps you recover it.

Comment

📞 Lost & Found Caller — CALL-E

The Problem

Losing something in a mall, airport, or public place is frustrating enough. Finding the lost-and-found department and repeatedly calling different places makes the process even harder.

We wanted to solve a simple question:

What if you could tell an AI what you lost, and the AI could make the phone call for you?

That idea became Lost & Found Caller, an AI-powered lost-and-found assistant built with CALL-E.

💡 What Inspired Us

We were inspired by a very ordinary real-world problem: when people lose something, the first step is often a phone call.

But making these calls can be repetitive:

  • Explain what was lost
  • Explain where and when it was lost
  • Ask whether the item was found
  • Describe the item in enough detail
  • Repeat the same information if contacting multiple locations

We realized that this is exactly the kind of task where a voice AI agent can be useful.

Instead of building another chatbot that simply answers questions, we wanted to build something that could take action in the real world.

🚀 What We Built

Lost & Found Caller lets a user provide information about their missing item and the location where they think they lost it.

The application then prepares the information needed for the call and uses CALL-E to communicate with the lost-and-found department.

The basic workflow is:

Describe → Call → Compare → Recover

The user doesn't have to manually make the call or repeatedly explain the situation.

CALL-E handles the conversation and brings the result back to the user.

⚙️ How It Works

1. Describe the lost item

The user enters information such as:

  • What they lost
  • When they lost it
  • A description of the item
  • Where they lost it

2. Select the location

The user chooses the relevant lost-and-found location.

3. CALL-E makes the call

The application sends the relevant information into the calling workflow.

CALL-E communicates with the lost-and-found department using a natural voice conversation.

Instead of simply generating text, the agent can actually make the phone interaction on the user's behalf.

4. Get a recovery signal

The conversation is converted into a useful result for the user, such as:

  • Likely Match
  • No Match
  • Uncertain

This turns a long phone conversation into a simple, actionable result.

🧠 Why CALL-E?

The key idea behind our project is that AI should not only answer — it should be able to act.

A traditional chatbot could tell a user:

"You should call the mall's lost-and-found department."

Our application takes the next step.

CALL-E actually makes the call.

That difference is what makes this project useful beyond a simple conversational interface.

🛠️ How We Built It

The application was built as a web application with a simple user interface for collecting lost-item information and initiating the calling workflow.

The main components are:

  • Frontend: Next.js / React
  • Voice AI: CALL-E
  • API routes: Next.js server-side API routes
  • Deployment: Vercel
  • Real phone interaction: CALL-E calling workflow

We also designed the application so that the user experience remains simple: the user provides the information once, and the agent handles the communication.

🔥 The Real-World Demo

The most important part of our demo is the real CALL-E call.

We first demonstrate the application and its simulated workflow.

Then we use the same information to initiate a real call.

This demonstrates the core idea of the project:

The AI isn't just responding on a screen — it is communicating with the real world.

🧩 Challenges We Faced

One of the biggest challenges was connecting a simple web interface with a workflow that involves an actual phone call.

We had to think carefully about:

  • How to collect the right information from the user
  • How to pass that information into the calling workflow
  • How to structure the call so the agent asks useful questions
  • How to turn a natural conversation into a simple result
  • How to handle situations where the lost-and-found department cannot confirm an item

Another challenge was making the experience feel like a real product rather than just an API demonstration.

We wanted the user journey to be extremely simple:

Enter the details → Let CALL-E call → Receive the result.

📚 What We Learned

Building this project taught us that voice AI becomes especially powerful when it is connected to a real task.

We learned how to:

  • Design an AI workflow around a real-world problem
  • Integrate a voice agent into a web application
  • Structure information for an AI phone conversation
  • Handle asynchronous calling workflows
  • Build a simple interface around a complex backend process
  • Think about AI as an agent that performs actions, rather than only a system that generates responses

Most importantly, we learned that a good AI application doesn't necessarily need to be complicated.

Sometimes the best use of AI is to remove one frustrating step from a person's everyday life.

🌎 What's Next?

Lost & Found Caller is just one example of what this type of AI agent can do.

The same architecture could eventually be extended to other tasks such as:

  • Calling businesses about forgotten items
  • Checking appointment availability
  • Following up on service requests
  • Calling customer-support departments
  • Making routine information calls

Our goal is to explore how voice agents can handle the repetitive phone interactions that people don't want to make themselves.

Lost something?

Let CALL-E make the call. 📞

Built With

Share this project:

Updates

Submission history