Caller AI — We wait. You talk.

Company: Caller AI — We wait. You talk.

What it appears to be: A proof-of-concept tool that uses AI to navigate phone IVR systems on behalf of users, then bridges them to a live agent when one becomes available. It is built as a demo for a hackathon and includes a public evaluator interface but no verified production use or revenue.

Key change: The project was submitted to the OpenAI 2026 hackathon and describes itself as an experiment in AI-driven phone navigation with strict authorization, deterministic execution, and limited data handling.

Single most important open question: Is there any evidence of traction, revenue, or customer adoption beyond the author's own demonstration?

The description states that Caller AI:

• Navigates a bounded IVR flow on behalf of a user.
• Waits in the queue until a live representative is ready.
• Calls the user back only when a live agent is available.
• Does not impersonate the customer, handle credentials, authorize purchases, negotiate, or continue speaking after the human joins.
• Uses GPT-5.6 to propose actions based on current IVR observation.
• Enforces deterministic policy validation before Twilio executes anything.
• Operates with strict JSON schema and user authorization.
• Includes a public demo that allows judges to see live GPT proposals without making real calls.

Inference: The product is an experimental system designed to reduce time spent on hold by automating IVR navigation, but it is not yet deployed in production for actual users.

The description states:

• The inspiration behind the project was that waiting on hold is a tax on people's time.
• The useful moment is the human conversation at the end—not navigating menus or listening to hold music.
• Caller AI gives that time back while keeping the customer in control.

Inference: The positioning is centered around user convenience and reducing friction in support interactions. It frames itself as a solution to a common pain point, but does not claim any market traction or adoption.

The description states:

• The system is designed for customers who interact with companies via phone support.
• Users must explicitly authorize a call and verify their callback number.
• A reviewed company route is required before dialing can occur.

Inference: The target customer appears to be individuals seeking help from businesses with IVR systems, but the product is not yet used by real customers. It is limited to authorized users in a controlled demo environment.

Not evidenced.

The description does not mention any pricing model, monetization strategy, or business model beyond the fact that it was built as a hackathon project.

The description states:

• Built with OpenAI, Twilio, and TypeScript.
• Each approved route is represented as a Call Graph.
• GPT-5.6 receives only current IVR observation and proposes one typed action under strict JSON Schema.
• Deterministic policy validates the proposal against graph, authorization, callback ownership, business hours, quotas, budget, and sensitive-data gates.
• Twilio Voice handles calls, DTMF, callbacks, and final human-to-human bridge.
• OpenAI API storage is disabled.
• Codex was a partner for implementation during Build Week expansion.

Inference: The technical architecture is designed with strong separation between AI interpretation and execution authority. It includes safeguards to prevent unauthorized actions or data leakage.

The description states:

• A private-pilot end-to-end UPS handoff with a verified callback and durable receipt.
• An independently enforced typed-action boundary.
• A live judge-facing GPT-5.6 evaluator with zero telecom side effects.
• Explicit refusal paths for credentials and unsupported requests.
• Meaningful post–July 13 Build Week extensions documented in the repository history.

Inference: There is evidence of development activity and some limited testing, but no verified customer base or revenue data. The system has not been deployed at scale or used by real customers.

Not evidenced.

The description does not mention competitors or similar products in the market.

• No production use: The system is described only as a hackathon demo with no verified live users.
• Limited scope: It works only within reviewed IVR flows and requires explicit user authorization.
• Unverified claims: All statements are self-reported and unverified; there is no evidence of traction, revenue, or customer adoption.
• No pricing or monetization strategy: No indication of how the product would be sold or funded in a real-world scenario.

1. What specific companies or departments have been reviewed for use with this system?
2. Has the system been tested with actual users beyond the demo environment?
3. Are there any plans to expand beyond the current demo scope, and if so, what are they?
4. How does the team intend to scale the number of supported IVR routes?
5. What is the plan for handling multilingual support or international use cases?