Description
AUI introduces Apollo-1, a groundbreaking neuro-symbolic foundation model designed for task-oriented dialogue. Unlike traditional language models that reason solely over open-ended language, Apollo-1 introduces a second path: reasoning in code, not weights. This approach allows generative AI to work effectively for users while neuro-symbolic AI works for companies by conversing with users and adhering to business rules.
Read More
Apollo-1 offers a unique set of capabilities that set it apart. It provides traceable decision-making, giving users a "white-box" view into the agent's reasoning process. The model is tool-native, capable of activating any number of tools, including APIs, CRMs, inventory systems, and other LLM agents, with 100% accuracy. This makes it adaptable for any company and any use-case.
Rule-based controllability is a key feature, allowing Apollo-1 to adhere to conditional rules, styles, preferences, and policies. This ensures unparalleled control over agent behavior. Furthermore, Apollo-1 delivers grounded responses, meaning agents respond with necessary context, sources, and tools, never relying solely on training data and always returning verifiable answers. The model also supports continuous fine-tuning, learning and improving with every interaction by automatically incorporating human feedback.
Apollo-1 is deployed across regulated and unregulated industries, serving as a foundation for agents that function on behalf of companies. It offers a guided experience, multiple tool integrations, robust guardrails, and ensures groundness in its responses through live integrations. This contrasts with LLM-based agents, where AUI's Apollo-1 aims to provide more predictable and reliable performance. The company has also announced a strategic partnership with Google Cloud, further solidifying its position in the AI landscape.
- CategoryAI Models & LLMs
- DeploymentCloud
- PlatformsWeb
AUI Apollo-1 Highlights
Neuro-symbolic reasoning in code, not weights
Traceable decision-making with a "white-box" view
Tool-native activation of APIs, CRMs, and LLM Agents
100% accuracy in tool activation
Rule-based controllability for adherence to policies and preferences
Grounded responses with context, sources, and tools
Continuous fine-tuning with human feedback
Supports regulated and unregulated industries
Foundation model for task-oriented dialogue
Predictable and reliable AI agent behavior
Getting Started with AUI Apollo-1
Explore Technology: Understand the core neuro-symbolic approach of Apollo-1.
Request API Access: Apply to integrate Apollo-1 into your systems.
Configure Agent Behavior: Define rules, styles, and policies for steerability.
Activate Tools: Connect Apollo-1 to necessary APIs, CRMs, and other systems.
Monitor Decisions: Utilize the white-box view for traceable reasoning.
Fine-tune Continuously: Allow the model to learn and improve from interactions.
AUI Apollo-1's Use Cases
- Task-Oriented Dialogue
- Rule-Based Automation
- Customer Service Agents
- Internal Business Tools
- Compliance Enforcement
