Cortex

What the company appears to be: Cortex is a self-reported AI-assisted outpatient workflow system designed for patients with brain tumours. The author describes it as a chat-based platform that uses an LLM to guide patient inquiries, triage urgency, and route cases to clinicians. It includes features like check-ins, automated advice, and clinician portals.

What changed: The project was built in the context of a hackathon (OpenAI 2026) and is described as a prototype with initial local evaluations underway. No commercial product or live deployment is evidenced.

Single most important open question: Is there sufficient clinical agreement and operational feasibility to support a scalable, real-world implementation? The description lacks evidence of actual user testing, clinician buy-in beyond the team, or any traction.

The description states that Cortex is an AI-assisted outpatient workflow system for patients with brain tumours. It includes:

• A patient-facing chat interface (progressive web app)
• LLM-driven triage and problem clarification
• Knowledge documents informing questions and red flag detection
• Automatic prioritization of urgency
• Alerting to clinical teams for urgent cases
• Check-in prompts via email
• Clinician portal with shared queue, history review, and endorsement/rejection workflows
• Policy document management and knowledge base updates
• Auditing capabilities for iterative improvement

The system is described as being built using ChatGPT 5.6 Sol and Codex, deployed on a VPS.

Evidence: Self-reported by the author. No independent verification or demonstration provided.

The description states that Cortex aims to formalize an existing outpatient pathway for brain tumour patients, improving experience for both patients and clinicians. It positions itself as addressing inefficiencies in current communication methods (e.g., email forwarding) and delays in response times.

It claims to improve patient outcomes by catching issues early through check-ins and ensuring timely escalation of urgent cases.

Inference: The project appears to be a proof-of-concept for AI-assisted clinical triage, likely targeting a niche but high-stakes area — oncology outpatient care. It is not positioned as a general-purpose tool or platform.

The description states that Cortex targets patients with brain tumours who are undergoing outpatient care, including those in diagnosis, counselling, surgery, and post-operative phases.

It also mentions clinicians managing these patients, particularly those involved in regular follow-up and unscheduled query handling.

Evidence: Self-reported. No evidence of customer segmentation or persona development beyond the stated clinical team and patient group.

The description does not provide any information about pricing, monetization, or business model.

Not evidenced.

The project was built using:

• ChatGPT 5.6 Sol for architecture design and implementation planning
• Codex for code generation
• Python as the development language
• Deployment on a VPS for initial testing

It is described as a progressive web app (PWA) for patients.

Evidence: Self-reported. No evidence of scalability, infrastructure, or production-grade delivery.

The description states that:

• The team is in the process of initial local evaluations
• They aim to work with a hospital trust for an early pilot
• It was built as part of a hackathon (OpenAI 2026)
• No revenue, customers, or adoption data are provided

Not evidenced.

The description does not mention any competitors or competitive landscape.

Not evidenced.

• Unverified clinical feasibility: The system is described as being built in collaboration with a clinical team but lacks evidence of broader clinical validation or regulatory alignment.
• Limited scope: Built for one specific use case (brain tumour outpatient care) and no indication of scalability to other conditions or workflows.
• No commercial traction: No customers, revenue, or product-market fit signals are evident beyond the hackathon prototype.
• Dependency on AI tools: Reliance on ChatGPT 5.6 Sol and Codex may not be sustainable or scalable in a clinical setting.
• Lack of operational evidence: No data on clinician adoption, patient feedback, or real-world performance.

1. What specific clinical workflows were identified and validated with your team?
2. How was the knowledge base for triage and advice developed? Who reviewed it?
3. Has any part of this system been tested in a real hospital setting, even informally?
4. What are the regulatory or compliance considerations for deploying such a system in a clinical environment?
5. How do you plan to scale beyond one team or one hospital trust?
6. Are there any existing tools or systems that this would integrate with or replace?
7. What is the timeline and budget for moving from pilot to full deployment?