TRI Evidence-to-Timeline Copilot

The description states that TRI is a web application designed to process fictional customer-service materials using AI and human review. It proposes structured timeline events derived from text inputs, each with:

• Source field
• Supporting evidence sentence
• Normalized time
• Time type (exact, approximate, range, unknown)
• Certainty level
• Human-review flag

A human reviewer can verify, correct, or approve these events. The system supports final reporting and audit records.

Inference The product appears to be a proof-of-concept prototype built for demonstration purposes rather than a production-ready tool.

The author states that TRI was created to turn invisible consumer time loss into an evidence-linked, human-reviewed timeline. It positions itself as a way to make time spent in service interactions measurable and accountable.

It claims to support a workflow where:

• AI proposes events
• Humans verify and correct
• Final timelines are approved

This reflects a human-in-the-loop design, emphasizing that AI supports judgment, not replaces responsibility.

Inference The positioning is aligned with accountability and transparency in service workflows. However, the claim lacks evidence of real-world deployment or impact beyond the hackathon context.

Not evidenced.

The description does not identify any specific customer segment or target persona beyond "consumers" in general. It refers to fictional case materials and does not describe actual users or organizations that would use this tool.

Not evidenced.

There is no mention of pricing, monetization strategy, or business model in the description. The project was built for a hackathon and is described as a prototype with no commercial traction.

The author states that TRI was built using:

• Frontend: Next.js, React, TypeScript
• Backend: Node.js, PostgreSQL, Prisma
• AI Tools: GPT-5.6, OpenAI Responses API, Structured Outputs
• Testing/Dev Tools: Vitest, Playwright, Vercel, Codex-assisted development

It includes features such as:

• Server-side validation and canonicalization
• Evidence linking
• Human review workflow
• Audit records
• Concurrency protection
• Safe logout
• Production end-to-end verification

Inference The technical stack suggests a modern web application with AI integration. However, no evidence of scaling, performance metrics, or production usage is provided.

Not evidenced.

The only demonstration mentioned involves:

• One fictional case submitted
• Two fictional text files included
• GPT initially proposed seven timeline events
• A human reviewer added one missing event
• Eight events were approved in total
• Zero review items remained

There is no evidence of real customers, revenue, or adoption beyond this single test case.

Not evidenced.

The description does not mention any competitors or existing solutions in the space. No market analysis or competitive positioning is provided.

• No real-world use: The entire demonstration uses fictional data and has no evidence of actual customer adoption.
• Prototype-only: Built for a hackathon; no indication of further development or production readiness.
• Unverified claims: All descriptions are self-reported, with no independent validation.
• Limited scope: The system is described only in the context of one test case, not general applicability.
• AI dependency without clarity on output quality or consistency: No data on accuracy or reliability beyond a single example.

1. What is the intended real-world use case for this tool?
2. Has there been any testing with actual users or service professionals?
3. Are there plans to expand beyond fictional cases into real customer data workflows?
4. How does the system handle edge cases or ambiguous inputs?
5. Is there a roadmap for scaling or commercializing this solution?