In this interview, Alden Do Rosario, CEO of Custom GPT, and Lee Gaul, a thought leader in the LLM and generative AI industries, discuss a recent benchmark where CustomGPT outperformed industry giants like OpenAI in answer accuracy. They explore the nuances of CustomGPT’s approach, particularly its focus on retrieval-augmented generation (RAG), which enhances the accuracy and relevancy of answers by utilizing an organization’s specific data rather than general internet knowledge.
Key insights from the discussion include:
- CustomGPT specializes in RAG, focusing on ensuring that AI-generated responses are accurate and based on the organization’s own data, which is crucial for businesses where precise information is critical.
- The company’s approach to hallucination (the AI generating inaccurate or fictitious information) is rigorous, with specific strategies to prevent it at each stage of the data processing pipeline, highlighting the importance of reliable AI responses in professional settings.
- CustomGPT’s system was independently benchmarked by Tonic.ai, confirming its superior performance in answer relevancy compared to other major AI platforms, demonstrating the effectiveness of their specialized focus on RAG for business applications.
