ConcourseGPT by CLEARPORT

The description states that ConcourseGPT converts airport concession RFPs, terminal drawings, public operating data, and award records into traceable opportunity intelligence. It allows users to:

• Rank relevant packages or units.
• Generate preliminary bid/no-bid recommendations.
• Explain primary reasons for scores.
• Open underlying unit and lease-outline drawings.
• Connect public sales, passenger, rent, investment, and area data.
• Generate editable pro formas.
• Recalculate financials as assumptions change.
• Flag missing information and conflicting evidence.
• Provide actionable next steps.
• Identify decisions requiring human approval.
• Expose an audit trace showing how recommendations were produced.
• Export evidence packages or executive pursuit briefs.

The system is described as not replacing staff but freeing up their time to submit better proposals. It uses GPT-5.6 for document interpretation and Codex for engineering tasks such as data extraction, normalization, mapping, and interface building.

Inference The product appears to be a proof-of-concept or prototype application built using AI tools (GPT and Codex), not a commercial SaaS offering with ongoing support or scalability features.

The author positions ConcourseGPT as a tool that addresses the opacity of airport concession procurement by turning fragmented data into actionable intelligence. The core claim is:

“Can the data ingestion capabilities of AI turn public airport-procurement data into a faster, more transparent pursuit decision while keeping necessary business judgments under human control?”

This thesis evolved from a personal experience leading an $180 million airport concessions program where operators struggled with high evaluation costs and lack of structured analysis.

The author also notes that the U.S. Department of Transportation’s October 2025 interim final rule disrupted longstanding partnership strategies, increasing uncertainty for both airports and operators—this context is used to justify the relevance of the tool.

Inference The positioning reflects a niche market need in the airport concessions industry, but it lacks evidence of adoption or traction beyond the author's own experience and prototype development.

The description identifies the primary user as airport concession operators, particularly small businesses seeking to enter the captive audience of airports. These are described as lacking time, resources, or domain expertise to evaluate RFPs effectively.

It also mentions that the tool supports different types of operators with varying brand portfolios, operating capabilities, opening capacity, and available capital—suggesting a need for operator-specific scoring logic.

Inference The target customer segment is likely small-to-medium-sized concessionaires in U.S. airports who are pursuing opportunities through competitive bidding processes. However, no explicit segmentation or persona details are provided beyond general industry roles.

There is no evidence of a business model or pricing structure in the description. The author states that ConcourseGPT is not designed to replace staff but to free up their time and energy. It is presented as a prototype/demo tool with no mention of monetization, licensing, subscriptions, or fees.

Inference No commercial model has been described or evidenced; this remains unknown.

The system was built using:

• GPT-5.6 for document interpretation and reasoning across long RFPs, exhibits, tables, addenda, financial requirements, and drawing context.
• Codex for engineering tasks including:
  • Traversing and organizing source folders
  • Extracting and normalizing records
  • Joining award results to units and drawings
  • Generating structured datasets
  • Implementing operator-specific scoring
  • Building editable pro formas
  • Creating terminal maps and workflows
  • Adding audit traces and visibility controls

The application is self-contained, runs in a standard browser (no API keys or external accounts required), and includes automated production checks.

Inference The technical stack suggests an AI-augmented prototype built with limited infrastructure dependencies. It does not appear to be a scalable SaaS platform but rather a demonstration tool for internal use or early-stage testing.

There is no evidence of traction, revenue, customers, or adoption beyond the author’s own development and demo cases. The project is described as a hackathon submission (Devpost entry) and includes three sample procurement cases:

• DFW RFP 0051324: 18 awarded packages, 38 offered spaces
• IAD Tier 2 East: 15 units
• LAX Terminal 5: 19 live units

These are used to demonstrate functionality but do not indicate real-world usage or impact.

Inference The product has not demonstrated measurable traction or maturity beyond a prototype-level demo. No evidence of ongoing development, user feedback loops, or commercial deployment exists.

The description does not provide any information about competitors or competitive landscape in the airport concessions intelligence space. It does not reference existing tools, platforms, or vendors that might address similar needs.

Inference There is no evidence of competitive analysis or awareness of existing solutions in this domain.

1. Unverified claims: All assertions about utility, traction, and impact are self-reported without corroboration.
2. No commercial model: No pricing, monetization strategy, or business model is described.
3. Prototype-only status: The tool is presented as a demo/prototype with no indication of production readiness or scalability.
4. Limited scope: Only three real-world cases are shown; no broader dataset or generalizability is demonstrated.
5. High uncertainty in data sources: Public records are noted to be incomplete or inconsistent, raising questions about reliability and accuracy.
6. Single-person team: The entire project was built by one individual (Faraji Robinson), suggesting limited capacity for scaling or iteration.

Inference The lack of external validation, traction, or commercial viability raises significant risk that the product may not meet market needs beyond its initial prototype phase.

1. What specific feedback have you received from concession operators or airport stakeholders?
2. How does ConcourseGPT handle discrepancies between different versions of public documents (e.g., addenda, updates)?
3. Are there any known limitations in how GPT-5.6 interprets complex RFP language or drawings?
4. Has the tool been tested with actual concessionaires or used in live bidding scenarios?
5. What are the technical and operational challenges in scaling this solution across multiple airports or regions?
6. How do you plan to validate or improve the accuracy of AI-generated recommendations over time?
7. Is there any intention to integrate real-time data feeds, such as live RFP announcements or updated wage/cost structures?
8. What is your roadmap for monetization and long-term product development?