CrunchLLM: Multitask LLMs for Structured Business Reasoning and Outcome Prediction

2025-09-12

AI Summary

Predicting startup exit success—via acquisition or IPO—remains a critical challenge in entrepreneurship research. Existing approaches struggle to effectively integrate structured data (e.g., funding rounds, industry tags from Crunchbase) with unstructured textual content (e.g., company descriptions, product summaries). This paper proposes an adaptive large language model (LLM) framework tailored to entrepreneurship, leveraging parameter-efficient fine-tuning and prompt engineering to jointly model structured features and textual semantics. Crucially, it incorporates an interpretable reasoning module to enhance decision transparency. Evaluated on a real-world startup dataset, the method achieves 82.3% accuracy—significantly outperforming conventional machine learning models and non-adapted baseline LLMs. The framework establishes a new paradigm for startup success prediction under high uncertainty, balancing predictive performance with human-understandable interpretability.

Technology Category

• Natural Language Processing: (Large) Language Models
• Machine Learning: Structured Learning
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Predicting the success of start-up companies, defined as achieving an exit through acquisition or IPO, is a critical problem in entrepreneurship and innovation research. Datasets such as Crunchbase provide both structured information (e.g., funding rounds, industries, investor networks) and unstructured text (e.g., company descriptions), but effectively leveraging this heterogeneous data for prediction remains challenging. Traditional machine learning approaches often rely only on structured features and achieve moderate accuracy, while large language models (LLMs) offer rich reasoning abilities but struggle to adapt directly to domain-specific business data. We present extbf{CrunchLLM}, a domain-adapted LLM framework for startup success prediction. CrunchLLM integrates structured company attributes with unstructured textual narratives and applies parameter-efficient fine-tuning strategies alongside prompt optimization to specialize foundation models for entrepreneurship data. Our approach achieves accuracy exceeding 80% on Crunchbase startup success prediction, significantly outperforming traditional classifiers and baseline LLMs. Beyond predictive performance, CrunchLLM provides interpretable reasoning traces that justify its predictions, enhancing transparency and trustworthiness for financial and policy decision makers. This work demonstrates how adapting LLMs with domain-aware fine-tuning and structured--unstructured data fusion can advance predictive modeling of entrepreneurial outcomes. CrunchLLM contributes a methodological framework and a practical tool for data-driven decision making in venture capital and innovation policy.

Problem

• Predicting startup success through acquisition or IPO outcomes
• Integrating structured business data with unstructured textual information
• Adapting large language models for domain-specific entrepreneurship prediction

Innovation

• Domain-adapted LLM with structured-unstructured data fusion
• Parameter-efficient fine-tuning with prompt optimization
• Interpretable reasoning traces for transparent predictions