AXIS | Preserving Human Meaning in AI Conversations
Research has moved into engineering. Engineering is moving toward application.
Most AI systems can answer a question.
The greater challenge is preserving what matters across an entire conversation.
AXIS is a model-agnostic coherence layer designed to help AI systems maintain human intention, recognize important changes, and reduce conversational drift.
Developed from Joe Trabocco's research into linguistic coherence and human-AI interaction, AXIS has progressed into a functioning system undergoing continued refinement and comparative evaluation.
Alongside it, Signal Reader examines conversations for changes in meaning, continuity, discovery, and safety, connecting its findings to specific conversational evidence.
Where We Stand
Our initial comparisons have produced encouraging results. We are now concentrating on measurement reliability, reproducibility, and safeguards as the technology approaches its first application.
Healthcare is our immediate priority.
Through Signal Literature | Health, co-founded by Joe Trabocco and Dr. Shivam Natarajan, our work focuses on a critical question:
What happens when an AI remembers a patient's words but loses their meaning?
Our goal is to make those failures observable, measurable, and ultimately less likely.
Public Demonstration
The AXIS public demonstration is temporarily offline while we refine the technology and its evaluation environment.
Internal testing continues. Independent clinical and performance validation is a forthcoming priority.
We look forward to sharing further results as the work advances.
The technology is functioning. The evaluation continues. The next step is proving its value beyond our own laboratory.
Explore Signal Literature | Health
AXIS and Signal Reader are research and development systems, not validated clinical decision-making tools.