Context-aware forecasting of inter-firm competitive dynamics with graph learning and multi-agent reasoning

Abstract The evolution of competitive dominance in technology-intensive industries is shaped by complex interactions among research activity, media visibility, strategic disclosures, public attention, and market signals. Understanding these dynamics is critical for strategic intelligence, yet existing predictive models often treat firms as isolated entities and provide limited support for contextualizing long-horizon firm-to-firm competitive forecasts. Here, we introduce an integrated framework that captures the relational nature of market competition by extending our prior Bayesian Multivariate Time-series Graph Neural Network with a newly designed, post hoc, lookahead-free multi-agent reasoning layer. The framework infers inter-firm relational dependencies endogenously from multimodal signals to forecast latent firm competitiveness and pairwise competitive gap trajectories with quantified predictive uncertainty. To evaluate strategic plausibility, specialized persona-driven language model agents critique the forecasted trajectories across technological, financial, and regulatory lenses using retrieved evidence. Evaluated under a rigorous date-based temporal protocol with strictly withheld test horizons, the framework consistently outperforms competitive structural baselines and yields well-calibrated prediction intervals. When applied to thirty leading AI-oriented technology firms, the framework uncovers asymmetric competitive momentum, contested market leadership, and emerging challenger dynamics. By bridging relational time-series forecasting with structured multi-agent deliberation, this study advances competitive dynamics analysis from purely accuracy-centric prediction toward context-aware, uncertainty-informed strategic intelligence.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-15
DOI
https://doi.org/10.1038/s41598-026-79480-4
Primary Topic
Competitive and Knowledge Intelligence
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Context-aware forecasting of inter-firm competitive dynamics with graph learning and multi-agent reasoning

Paul D. Yoo, Daeseung Park, Seongil Han, Yonghwan Kwon
Scientific Reports
Competitive and Knowledge Intelligence
article

Context-aware forecasting of inter-firm competitive dynamics with graph learning and multi-agent reasoning

Paul D. Yoo, Daeseung Park, Seongil Han, Yonghwan Kwon
article en

Abstract

Abstract The evolution of competitive dominance in technology-intensive industries is shaped by complex interactions among research activity, media visibility, strategic disclosures, public attention, and market signals. Understanding these dynamics is critical for strategic intelligence, yet existing predictive models often treat firms as isolated entities and provide limited support for contextualizing long-horizon firm-to-firm competitive forecasts. Here, we introduce an integrated framework that captures the relational nature of market competition by extending our prior Bayesian Multivariate Time-series Graph Neural Network with a newly designed, post hoc, lookahead-free multi-agent reasoning layer. The framework infers inter-firm relational dependencies endogenously from multimodal signals to forecast latent firm competitiveness and pairwise competitive gap trajectories with quantified predictive uncertainty. To evaluate strategic plausibility, specialized persona-driven language model agents critique the forecasted trajectories across technological, financial, and regulatory lenses using retrieved evidence. Evaluated under a rigorous date-based temporal protocol with strictly withheld test horizons, the framework consistently outperforms competitive structural baselines and yields well-calibrated prediction intervals. When applied to thirty leading AI-oriented technology firms, the framework uncovers asymmetric competitive momentum, contested market leadership, and emerging challenger dynamics. By bridging relational time-series forecasting with structured multi-agent deliberation, this study advances competitive dynamics analysis from purely accuracy-centric prediction toward context-aware, uncertainty-informed strategic intelligence.

Scientific Reports
University of Suwon (KR), Birkbeck, University of London (GB)
Competitive and Knowledge Intelligence
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon