Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

P1GPT: a multi-agent LLM workflow module for multi-modal financial information analysis

Chen-Che Lu, Yun-Cheng Chou, Teng-Ruei Chen

TL;DR

P1GPT presents a layered multi-agent LLM workflow for multi-modal financial analysis and trading decision support. It implements a five-layer architecture (Input, Planning, Analysis, Integration, Decision) with Intelligent Specialized Agents and Supporting Agents, plus an integration-layer to fuse cross-modal outputs with transparent rationales. Backtests on AAPL, GOOGL, and TSLA show superior cumulative and risk-adjusted performance relative to baselines, along with interpretable, case-level explanations. The work argues that structured reasoning workflows, rather than mere role imitation, deliver scalable, explainable, and trustworthy financial AI suitable for deployment.

Abstract

Recent advances in large language models (LLMs) have enabled multi-agent reasoning systems capable of collaborative decision-making. However, in financial analysis, most frameworks remain narrowly focused on either isolated single-agent predictors or loosely connected analyst ensembles, and they lack a coherent reasoning workflow that unifies diverse data modalities. We introduce P1GPT, a layered multi-agent LLM framework for multi-modal financial information analysis and interpretable trading decision support. Unlike prior systems that emulate trading teams through role simulation, P1GPT implements a structured reasoning pipeline that systematically fuses technical, fundamental, and news-based insights through coordinated agent communication and integration-time synthesis. Backtesting on multi-modal datasets across major U.S. equities demonstrates that P1GPT achieves superior cumulative and risk-adjusted returns, maintains low drawdowns, and provides transparent causal rationales. These findings suggest that structured reasoning workflows, rather than agent role imitation, offer a scalable path toward explainable and trustworthy financial AI systems.

P1GPT: a multi-agent LLM workflow module for multi-modal financial information analysis

TL;DR

P1GPT presents a layered multi-agent LLM workflow for multi-modal financial analysis and trading decision support. It implements a five-layer architecture (Input, Planning, Analysis, Integration, Decision) with Intelligent Specialized Agents and Supporting Agents, plus an integration-layer to fuse cross-modal outputs with transparent rationales. Backtests on AAPL, GOOGL, and TSLA show superior cumulative and risk-adjusted performance relative to baselines, along with interpretable, case-level explanations. The work argues that structured reasoning workflows, rather than mere role imitation, deliver scalable, explainable, and trustworthy financial AI suitable for deployment.

Abstract

Recent advances in large language models (LLMs) have enabled multi-agent reasoning systems capable of collaborative decision-making. However, in financial analysis, most frameworks remain narrowly focused on either isolated single-agent predictors or loosely connected analyst ensembles, and they lack a coherent reasoning workflow that unifies diverse data modalities. We introduce P1GPT, a layered multi-agent LLM framework for multi-modal financial information analysis and interpretable trading decision support. Unlike prior systems that emulate trading teams through role simulation, P1GPT implements a structured reasoning pipeline that systematically fuses technical, fundamental, and news-based insights through coordinated agent communication and integration-time synthesis. Backtesting on multi-modal datasets across major U.S. equities demonstrates that P1GPT achieves superior cumulative and risk-adjusted returns, maintains low drawdowns, and provides transparent causal rationales. These findings suggest that structured reasoning workflows, rather than agent role imitation, offer a scalable path toward explainable and trustworthy financial AI systems.
Paper Structure (36 sections, 7 figures, 2 tables)

This paper contains 36 sections, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background and Motivation
  3. Challenges in Financial Trading
  4. Contributions of P1GPT
  5. Related Work
  6. Financial LLMs and Automated Assistants
  7. LLMs as Autonomous Traders
  8. Multi-Agent Systems in Financial Applications
  9. P1GPT System Overview
  10. Overall System Architecture
  11. Data Sources and Collection Pipeline
  12. Core Modules and Agent Roles
  13. Intelligent Specialized Agents (ISA)
  14. Supporting Agents
  15. Integration Strategy and Data Fusion
  16. ...and 21 more sections

Figures (7)

  • Figure 1: P1GPT overall workflow. Left: Input and Planning transform user queries and raw multi-modal data into an executable plan. Center: the Analysis Layer hosts specialized agents—Controller, Fundamental ISA, Tech ISA, Chip (Semiconductor) ISA, News ISA—and supporting modules (Search, Revenue, Trend Analyzer, Recommender). Right: the Integration Layer consolidates structured reports and forwards them to the Decision Layer, which outputs an Action with an explanation. Arrows denote data/control flow and standardized report interfaces between layers.
  • Figure 2: Trading trajectory of P1GPT on AAPL (Feb. 3–Sep. 30, 2025), showing model-generated buy (blue) and sell (orange) signals, along with position and cumulative return evolution.
  • Figure 3: Trading trajectory of P1GPT on GOOGL (Feb. 3–Sep. 30, 2025), showing model-generated buy (blue) and sell (orange) signals, along with position and cumulative return evolution.
  • Figure 4: Trading trajectory of P1GPT on TSLA (Feb. 3–Sep. 30, 2025), showing model-generated buy (blue) and sell (orange) signals, along with position and cumulative return evolution.
  • Figure 5: Cumulative Returns on AAPL using P1GPT. The figure shows the performance comparison of our model against baseline approaches for Apple Inc. stock analysis.
  • ...and 2 more figures