Casebook AI

The description states:

• Casebook AI is a local-first application built in Go with an embedded HTML/CSS/JavaScript interface and SQLite database.
• It supports compliance investigations, where users paste reports, policies, and evidence into a case.
• An AI model (GPT-5.6-sol) generates a structured brief including allegations, timeline, policy matches, gaps, risk flags, and recommended actions.
• Each factual claim includes a source quote that is independently verified against the submitted text.
• The system enforces human review of all AI-generated items; no silent edits or final determinations are made by the model.
• All decisions and actions are appended to an audit trail and can be tracked in checklist format.

This is a structured, human-controlled AI workflow tool for regulated environments, not a general-purpose AI assistant or document summarizer.

The description states:

• The product was originally built as a local case-management tool for small compliance teams.
• It was extended during OpenAI Build Week to include AI assistance that focuses on the hardest part of an investigation—generating structured, sourced summaries without pretending the model is the decision-maker.
• The author emphasizes that the AI output is operational, not decorative: accepted actions become trackable work items.
• The core value proposition is human control and auditability, not automation or speed.

Inference The positioning has evolved from a simple local tool to a human-in-the-loop AI workflow tailored for compliance environments, with an emphasis on trust, transparency, and defensibility.

The description states:

• The product is designed for small compliance teams, who are likely working in regulated industries such as finance, healthcare, or legal.
• It supports compliance investigations that involve messy data, deadlines, and need for audit trails.
• The author notes that the tool was built for a specific use case: “a reviewer may receive a report, policy passages, email excerpts, a partial system export, and a deadline.”

Inference The ICP likely includes compliance officers, legal investigators, or auditors in regulated industries who need to manage complex, evidence-based investigations with strict provenance requirements.

Not evidenced.

The description does not mention any pricing model, monetization strategy, or business model. It only describes the tool’s functionality and design principles.

The description states:

• The application is built in Go, with embedded HTML/CSS/JavaScript UI and a local SQLite database.
• It uses the OpenAI Responses API with GPT-5.6-sol, strict JSON Schema output, and input limits.
• Safeguards include:
  • Explicit user submission notice
  • store: false
  • Input limits and SHA-256 packet fingerprinting
  • Per-user hashed safety identifier
  • Exact-quote grounding post-generation
  • Owner/coverer permission checks
  • Immutable, idempotent review decisions
  • Append-only provenance events
• A deterministic synthetic fixture allows testing without an API key.

Inference The tool is designed with privacy and auditability as core principles. It uses a local-first architecture, which suggests it may be suitable for environments where data sovereignty is critical.

Not evidenced.

There is no mention of customers, revenue, usage metrics, or adoption data. The description only details the tool’s design and functionality.

Not evidenced.

The description does not reference any competitors or market landscape. It does not describe how Casebook AI compares to existing tools in compliance case management or AI-assisted investigation platforms.

• Single-person team: The project is built by one person (1/10th of a typical M&A due diligence team).
• No traction or customers: No evidence of revenue, users, or adoption beyond the author’s own use case.
• Niche application: The tool is tailored for compliance investigations in regulated environments—this may limit its scalability or market size.
• AI model dependency: Reliance on GPT-5.6-sol and OpenAI API introduces risk of API availability, cost, or model changes.
• Local-first architecture: While privacy-focused, this may limit enterprise adoption unless a cloud version is planned.

1. What specific compliance industries or use cases are you targeting?
2. How do you plan to scale beyond a single developer and local-first architecture?
3. Are there any existing customers or pilot users of the tool?
4. What is your roadmap for enterprise features like authentication, TLS, role-based access control, and encryption?
5. How do you intend to monetize this product, if at all?
6. What are the technical challenges in moving from a local-first model to a cloud-hosted version?
7. Are there any regulatory or compliance concerns around AI use in your target domain that could affect adoption?