The Meeting Room

The Meeting Room is a self-reported tool built by a single founder (Doddy Samiaji), an architect/urban designer, for use in professional decision-making processes involving AI reviewers shaped as credible experts. The product is described as part of a suite of tools developed at Kolabs.Design, aimed at enabling users to assemble and manage AI agents representing different professional roles—such as planners, engineers, lawyers, or investors—to evaluate proposals, strategies, or policies.

The author states that the tool supports complex decision-making by allowing independent review from multiple AI experts, followed by a synthesis report that includes risks, disagreements, and next steps. It uses technologies like OpenAI Codex 5.6, Next.js, Vercel AI SDK, and others to implement its functionality.

No revenue, customers, or traction data are provided beyond the self-reported description. The tool is not yet available for public use, nor does it appear to have been deployed in production. The project was submitted to the OpenAI 2026 hackathon on Devpost.

The description states that The Meeting Room is a tool for assembling and managing AI reviewers who represent different professional roles (e.g., financial adviser, planner, engineer, lawyer) in order to evaluate proposals or decisions. Each reviewer has a defined background, responsibility, and question.

It allows users to:

• Bring in a proposal or decision case.
• Assemble two to eight AI reviewers shaped as credible professionals.
• Assign distinct tasks to each reviewer without them seeing others' answers (to prevent influence).
• Optionally engage a "Deep Review" where reviewers can reconsider their positions based on anonymous summaries of other views.
• Download a final PDF report summarizing agreement, disagreement, risks, assumptions, and next steps.

The human principal always makes the final decision.

Inference The tool appears to be an experimental interface for structured AI-assisted expert review, not a general-purpose AI assistant or decision engine.

The author claims that The Meeting Room is designed to simulate a "serious professional review" — one that includes independent perspectives, constructive disagreement, and clear synthesis. It is positioned as an alternative to consulting a single AI, which the author describes as feeling like “consulting one intelligent person.”

The tool is said to reflect how real professional decisions are made, incorporating multiple viewpoints without forcing artificial consensus.

Inference The positioning evolves from a personal need (to improve decision-making in design and planning) into a broader claim about how AI can be used to support collective intelligence in complex domains.

The description states that the tool is being built for use by professionals working in fields such as architecture, urban planning, design, development, and policy. These users are described as those who bring together diverse stakeholders (designers, engineers, investors, lawyers, government officials) to make decisions.

It is implied that the primary user is a principal decision-maker—such as an architect or planner—who wants to evaluate complex cases using expert perspectives from AI agents.

The tool is also said to be part of a suite being developed at Kolabs.Design, suggesting internal use within a practice rather than direct market targeting.

Inference The ICP likely includes professionals in design, planning, and policy-making who seek structured expert input but may not have access to full teams of specialists.

There is no evidence provided regarding pricing, monetization, or business model. The description does not mention any revenue streams, customer acquisition plans, or commercial partnerships.

The tool is described as one of several being built at Kolabs.Design and is intended for internal use.

Inference No clear indication of how the product will generate value or be sold; it may be an experimental prototype or a tool for internal practice use.

The project was built using:

• Technologies: clerk, codex, drizzle-orm, google-stitch, gpt5.6, neon-postgres, next.js, openai, openai-api, openai-codex, pdf-lib, react, tailwind-css, typescript, vercel-ai-sdk, vercel-blob, vercel-workflow, zod
• Framework: OpenAI Codex 5.6 was used to translate domain knowledge into code

The author emphasizes that they did not become a software engineer but instead focused on defining how the process should work, while Codex supported technical implementation.

Inference The tool is built using modern web and AI stacks, with a focus on integrating AI models for expert simulation. However, no evidence of deployment or scalability beyond prototype status.

There is no evidence of traction, adoption, or usage metrics. The project is described as being in development, part of a suite at Kolabs.Design, and submitted to a hackathon.

The author mentions testing it on real cases and improving it through daily use, but does not provide data on how many cases have been reviewed or what the outcomes were.

Inference The tool has not yet reached a production stage or demonstrated measurable impact in practice. It is likely an early-stage prototype.

The description does not mention specific competitors or similar tools. However, it implies that the product addresses a gap in current AI decision-support systems by enabling structured expert review and disagreement among AI agents.

It is positioned as a way to avoid echo chambers and ensure independent evaluation — features that are not commonly found in single-AI decision tools.

Inference The competitive landscape includes general-purpose AI assistants, but there is no clear evidence of direct competitors or market positioning against them.

• Lack of traction or usage data: No evidence of real-world application or user feedback.
• Single-founder development: Only one person involved in building the tool; unclear if it has been tested beyond internal use.
• Unproven utility: The author claims the approach improves decision-making, but no validation is provided.
• No commercial viability: No pricing, monetization, or customer model described.
• Prototype nature: Submitted to a hackathon and not yet in production; unclear if it will evolve into a product.

Inference The tool lacks evidence of market demand or proven effectiveness. It remains an experimental idea with no demonstrated path to commercialization.

1. What specific problems are you trying to solve, and how do you know they exist?
2. Have you tested this approach on real cases? If so, what were the results?
3. How does this differ from existing AI tools or expert review processes?
4. Is there a plan for scaling beyond internal use at Kolabs.Design?
5. What is your long-term vision for monetizing or deploying The Meeting Room?
6. How do you ensure that AI-generated reviews are not misleading or biased?