Artificial intelligence has become increasingly adept at mimicking human conversation and dispensing advice, but significant concerns remain about accuracy, privacy, and fair compensation for creators. A newly launched startup called Onix is attempting to address these challenges by creating AI chatbots trained directly by experts in their respective fields, allowing subscribers to access personalized guidance whenever they need it.
Startup Addresses AI Accuracy and Creator Compensation
Founded by David Bennahum, Onix positions itself as a platform where experts can create AI versions of themselves that conduct conversations with paying subscribers. The concept draws parallels to subscription-based creator platforms, where users pay to access specialized knowledge and advice from trusted professionals. The chatbots are designed to replicate the conversational style and expertise of their human counterparts, delivering consultations that feel personalized and informed by years of professional experience.
Expert-Trained Chatbots with Privacy Protections
The company emphasizes privacy and data security as core differentiators. User information is stored on individual devices with encryption, meaning the Canada-based company retains minimal personal data even if legally compelled to share it. Since experts train the models using their own content, Onix argues there are no intellectual property concerns—creators maintain ownership of their knowledge while monetizing it directly.
Limitations in Domain Constraints and Guardrails
To minimize hallucinations, Onix implements guardrails that constrain conversations to relevant topics. However, early testing reveals these safeguards have limitations. When prompted to discuss subjects outside their trained domain, the bots occasionally generate false information rather than decline engagement.
Emerging Competition in Expert AI Services
The platform is currently in beta, accepting testers from a waitlist before opening to broader audiences. While Onix presents itself as innovative, similar services already exist. Parenting expert Becky Kennedy has successfully built a business around an AI chatbot named Gigi trained on her knowledge, reportedly raising $34 million in funding.
As AI-powered expert consultation services proliferate, the model raises important questions about the future of professional advice, the value of human expertise in an AI-augmented world, and how revenues are distributed between platforms and the experts whose knowledge trains these systems.