OpenAI 2026 hackathon

Delegate

An AI meeting agent you can actually trust. It only answers with evidence, defers to you when needed, and can screen share to clients – with barely any conversational latency.

Solo project by Ray Sankhyan · 0 likes · 0 comments

Archive position — measured, not model output

0 likes on Devpost

2,264 of the 7,856 archived projects have more likes, and 5,592 share exactly 0 — so this project's #3,694 place in the like-ranked listing is a tie-break inside that group, not a ranking.

Projects (log scale)

1
10
100
1k
10k
05,592
11,758
2285
3–4132
5–975
10+14

Likes on Devpost. ▲ marks this project's group.

Show the figures
LikesProjectsShare of archive
05,59271.2%
11,75822.4%
22853.6%
3–41321.7%
5–9751.0%
10+140.2%
Devpost like counts for all 7,856 archived projects, captured when this archive was built.

Executive Summary

The company appears to be a solo-built AI meeting agent named Delegate, designed to represent users in Zoom meetings with evidence-based responses, live browser presentation, and authority-aware decision-making. The author states that it uses GPT-5.6 models, speech-to-text, text-to-speech, and browser automation tools to join meetings, retrieve relevant documents, and present information visually.

What changed: This project is a self-reported hackathon submission, not a commercial product or service. It represents an experimental approach to AI meeting agents that go beyond passive summarization toward active representation with guardrails.

Single most important open question: Is there any evidence of actual user adoption, revenue, or traction beyond the author's own description?

Back to contents

What The Product Actually Is

The description states that Delegate is an AI meeting agent that:

  • Joins Zoom meetings in real time
  • Listens and speaks during meetings
  • Represents the user’s position while they are away
  • Retrieves evidence from documents before answering questions
  • Presents webpages or workflows live during meetings
  • Makes decisions based on a defined "meeting brief" with goals, policies, and reference materials
  • Logs all actions in a post-meeting commitment ledger

It is built using JavaScript/Node.js and integrates tools like GPT-5.6 Terra/Luna, Deepgram Flux, Attendee.dev, Browserbase, and OpenAI's text-embedding-3-small.

Not evidenced: No information about actual functionality, performance, or user experience beyond the author’s claims.

Back to contents

Positioning & Claim Evolution

The author states that Delegate was built to solve a single bottleneck in enterprise AI adoption: "What would it take to actually trust an AI to represent you in your meetings?"

It positions itself as:

  • A real-time meeting representative, not just a passive recorder or summarizer
  • An agent that answers with evidence, defers when needed, and can screen share
  • A tool that bridges verbal explanation and visual demonstration
  • A trustworthy alternative to generic AI tools

The author also says it is exploring a new category: moving away from note-taking toward autonomous, responsible AI representation.

Inference: The positioning implies an attempt to differentiate from existing AI meeting tools by emphasizing trustworthiness, evidence-based responses, and visual demonstration capabilities.

Not evidenced: No market research, competitive analysis, or user feedback on positioning.

Back to contents

Target Customer & ICP

The description states that Delegate is built for enterprise users, particularly those seeking to:

  • Represent themselves in meetings without being present
  • Ensure AI responses are grounded in policy and documentation
  • Have a tool that knows when to escalate decisions back to the human owner
  • Use AI to demonstrate workflows or information visually during meetings

It is implied that enterprise decision-makers who rely on structured processes, compliance, and accountability would be the primary users.

Not evidenced: No specific customer segments, personas, or use cases beyond the author’s own vision.

Back to contents

Business Model & Pricing Evidence

The description does not contain any information about:

  • How the product will be monetized
  • Whether it is a SaaS offering or a one-time tool
  • What pricing structure might exist (if any)
  • Any revenue model, subscription plans, or licensing terms

Not evidenced: No business model or pricing details.

Back to contents

Technical & Delivery Signals

The author states that Delegate was built using Codex, with GPT-5.6 Terra and Luna models, JavaScript/Node.js, and a tech stack including:

  • Deepgram Flux (speech recognition/synthesis)
  • Attendee.dev (Zoom integration)
  • Browserbase (virtual browser control)
  • OpenAI's text-embedding-3-small (document grounding)

The system is described as using a state machine to coordinate audio streams, browser actions, and user dashboards.

Not evidenced: No information on scalability, reliability, or production deployment. No mention of infrastructure, data handling, or security practices.

Back to contents

Traction & Maturity Signals

The description states that this is a hackathon submission, not a commercial product.

It includes no evidence of:

  • Customers or users
  • Revenue or monetization
  • Product usage metrics
  • Iteration history or versioning
  • Any form of market validation or feedback

Not evidenced: No traction, adoption, or maturity beyond the initial build.

Back to contents

Competitive Context

The description does not include any information about:

  • Competitors in the AI meeting agent space
  • How Delegate compares to existing tools (e.g., Otter.ai, Notion AI, Zoom’s AI features)
  • Market size or competitive positioning

Not evidenced: No competitive analysis or market context.

Back to contents

Key Risks & Red Flags

The author states that Delegate is a solo-built hackathon project, with no team beyond one person (Ray Sankhyan).

Key risks and red flags include:

  • Single-person development: No team, no support structure, no scalability
  • Unverified claims: All features are self-reported without independent validation
  • No traction or revenue: The tool is not yet commercialized or adopted
  • Unclear path to market: No evidence of product-market fit or go-to-market strategy
  • High technical complexity: Integrating Zoom, browser automation, and AI reasoning in real time is technically challenging

Inference: The project may be a proof-of-concept rather than a viable commercial offering.

Back to contents

Diligence Questions To Ask The Founders

  1. What specific enterprise use cases have you validated with potential users?
  2. How do you plan to scale beyond a single developer’s capacity?
  3. Have you tested the system in live meetings or only in simulation?
  4. What is your roadmap for integrating calendar tools, organization-wide policies, and other platforms?
  5. Are there any legal or compliance concerns around AI representation in meetings?
  6. How do you intend to monetize this tool if it’s not a SaaS product?

Back to contents

Investment/Partnership Verdict

The description states that this is a hackathon submission, not a commercial venture.

There is no evidence of:

  • Revenue, customers, or traction
  • A clear business model or pricing strategy
  • Team size beyond one person
  • Product maturity or scalability

Verdict: This is an experimental idea with strong technical ambition but no demonstrated commercial viability. It is not ready for investment or partnership at this stage.

Confidence level: Low — based entirely on self-reported, unverified information.

Back to contents

Source

Submitted to the OpenAI 2026 hackathon on Devpost. Project home on DevPost.

The analysis above was generated by a language model from the project's own one-line description. It is not independent research and contains no verified traction, revenue or customer data.