AI agents build AI agents in new OASYS platform

Key Concepts

• OASIS (Orchestrated Agent System): A self-learning, agentic AI platform that allows businesses to build and deploy custom conversational AI agents.
• Agentic AI: AI systems capable of performing tasks, learning, and adapting autonomously rather than just responding to static prompts.
• Edge AI: The deployment of AI models directly on local devices (cars, robots, IoT devices) rather than relying solely on cloud-based data centers.
• Voice AI: Specialized AI technology focused on speech recognition and natural language processing for audio-based interactions.

---

1. Main Topics and Key Points

The video features an interview with the CEO of SoundHound AI regarding their new platform, OASIS. The core value proposition is the transition from manual, human-heavy workflows to automated, self-improving AI agents.

• Efficiency: Tasks that previously required months of development and large teams can now be completed in minutes.
• Scalability: The platform is designed to handle high-volume environments, such as call centers with 10,000+ operators, by automating customer interactions.
• Self-Learning: Once deployed, agents continuously learn from interactions and suggest improvements to human supervisors for final approval, ensuring the system remains accurate and "on-rails."

2. Real-World Applications

• Customer Service/Call Centers: Automating high-volume inbound calls for banks and insurance companies to reduce wait times and operational costs.
• Food Service: Implementation in drive-thrus (e.g., White Castle) to handle complex orders, such as specific dietary modifications (e.g., "no pickles").
• Hospitality: Managing reservations and service requests at locations like Resorts World, Las Vegas.
• Edge Computing: Integrating AI into hardware like cars, robots, and IoT devices to function without a constant cloud connection.

3. Methodology: The OASIS Framework

The process for deploying an AI agent via OASIS follows these steps:

1. Data Ingestion: The business uploads existing conversation data (e.g., call logs) into the platform.
2. Description: The user describes the desired task or objective for the AI.
3. Automated Building: The OASIS platform uses its own AI to build the specific agent required for the task.
4. Deployment & Learning: The agent goes live, processes queries, and learns from real-world interactions.
5. Human-in-the-Loop: The system presents recommended improvements to a human supervisor, who reviews and approves changes to prevent the AI from deviating from business standards.

4. Key Arguments and Evidence

• Revenue Growth: The CEO cited a study from a drive-thru customer showing that locations using SoundHound’s AI experienced higher revenue compared to locations without it, suggesting that AI improves both efficiency and sales performance.
• Accessibility: The platform supports multi-language capabilities, allowing businesses to serve a global customer base without language barriers.
• Cost Reduction: By automating call centers—often the most expensive cost center for large enterprises—businesses can significantly reduce overhead while eliminating long wait times for customers.

5. Technical Concepts

• Speech Recognition Engine: SoundHound’s proprietary technology that powers the audio-to-text and intent-understanding capabilities.
• Foundation Models: The underlying AI architectures that allow the agents to understand and generate human-like responses.
• Edge vs. Cloud: The CEO emphasizes that while most complex AI runs in the cloud (requiring expensive GPUs and data centers), SoundHound’s technology allows for "Edge" deployment, which is more efficient for hardware like vehicles and robots.

6. Synthesis and Conclusion

SoundHound AI’s OASIS platform represents a shift toward "AI building AI." By focusing on voice-based, agentic systems that can operate at the "edge," the company aims to solve the scalability issues faced by large enterprises. The primary takeaways are that AI is moving beyond simple chatbots to autonomous agents that not only reduce operational costs but also drive measurable revenue growth. The company’s strategy involves deep collaboration with chipmakers to optimize performance on local hardware, positioning them to serve a wide range of industries from banking to automotive.