PELICAN

Pelican AI: The Specialized Research Interface

Standard AI models are limited by "knowledge cutoffs" and a tendency to hallucinate when analyzing complex financial data. Pelican AI is a specialized orchestration engine built to solve these problems. It connects our proprietary investment research, raw regulatory filings, and real-time market data into a single conversational interface.

Accuracy Through Precision Retrieval

Our agents do not "guess" or rely on general knowledge. Instead, they use a process called Contextual Retrieval. When you ask a question, the engine identifies the specific source required—whether it’s a paragraph in a 10-K, a line from an earnings call, or a real-time trading multiple—and "snips" that exact data for the LLM to review. This ensures every answer is grounded in verifiable fact.

Capabilities of the Agent Layer

You don’t need to know which agent to call; our interface automatically routes your request to the specialist best equipped for the task.

1. Investment Thesis & Synthesis

This capability serves as the authority on Pelican’s completed research. It synthesizes our core investment theses, connecting structural business quality to specific catalysts and risks.

  • Example: "Summarize the core investment thesis and primary risks for NVIDIA."

2. Deep Filing Analysis (10-K, 20-F & Proxy)

These specialists are trained to navigate the nuances of SEC filings. They can extract granular data on revenue segments, customer concentration, risk factors, or executive compensation structures that are often missed by standard data aggregators.

  • Example: "What are the specific vesting conditions for the CEO’s performance-based stock units?"

3. Live Market & Valuation Intelligence

Unlike standard LLMs, this agent has no knowledge cutoff. It queries external financial APIs in real-time to retrieve current pricing, analyst estimates, and the latest trading multiples (EV/EBITDA, P/E, EV/Sales). It also monitors the tape for fresh 8-K and 6-K filings as they are released.

  • Example: "What is the current EV/Sales multiple based on today's price and TTM revenue?"

4. Narrative Tracking (Earnings Call Transcripts)

This specialist focuses on the "vibe and the numbers" from quarterly calls. It analyzes transcripts across multiple years to track management’s consistency, tone, and responses to analyst skepticism during Q&A sessions.

  • Example: "How has management’s commentary on international expansion changed over the last four calls?"

5. Idea Generation & Institutional Flow

This agent identifies new investment opportunities by searching through various "signal" feeds, including institutional accumulation (13Fs), insider trading activity, and specific quantitative screens.

  • Example: "Which high-conviction hedge funds increased their positions in healthcare stocks last quarter?"

Auditable Evidence: The Traceable Chain

In professional finance, a summary is only useful if it is auditable. Every response generated by Pelican AI includes Inline Citations that link directly to the source material.

  • Research Citations: (Investment Thesis | NVIDIA | 2025-01-10)
  • Regulatory Citations: (10-K | FY2024 | Item 1A. Risk Factors)

This allows you to instantly verify the logic and data behind every AI-generated statement.

Strategic Use Cases

  • Validate Assumptions: Cross-reference a management claim against the "Risk Factors" section of their latest filing.
  • Monitor the Tape: Instantly summarize a material event filing (8-K) released minutes ago.
  • Deep-Dive Research: Pull a comprehensive breakdown of revenue streams and growth drivers without manually parsing a 100-page document.
  • Contextual Sizing: Evaluate a stock’s current valuation relative to its historical trading anchors and peers.