SBIR Phase I: Revolutionizing Enterprise Processes with Automated Logical Reasoning Across Unstructured Data

About This Grant

The broader/commercial impact of this Small Business Innovation Research (SBIR) Phase I project is to transform enterprise knowledge work by improving how businesses analyze and act on unstructured data. Many industries such as technology, finance, and healthcare, rely on knowledge workers to extract insights from vast amounts of unstructured information, including documents, emails, and reports. However, existing AI solutions often produce unreliable or inconsistent results, limiting their effectiveness in high-stakes environments. This project introduces a hybrid approach that combines structured algorithmic workflows with advanced AI models, ensuring complex multi-step tasks are completed with high accuracy, transparency, and explainability. Initially targeting enterprise sales, where data-driven insights fuel revenue growth, this innovation will assist in researching customers, uncovering new opportunities, and streamlining deal-making processes. The technology provides a durable competitive advantage by providing enterprise level automations with high reliability and transparency, allowing businesses to trust and integrate AI in their processes. As a key enabler of commercial success, it positions the company as a leader in enterprise AI. By year three, this technology is projected to impact over one million knowledge workers and drive measurable gains in productivity and revenue generation. This Small Business Innovation Research (SBIR) Phase I project aims to develop a novel AI framework that integrates algorithm design (logical reasoning) with machine learning models to enhance the analysis of unstructured data. Unlike conventional AI approaches that rely solely on large language models (LLMs), this framework structures AI workflows as step-by-step processes, selectively incorporating models like LLMs where appropriate to ensure transparency, consistency, and accuracy. The research will investigate key questions, including: Can a structured, logic driven AI system outperform end-to-end LLM-based methods in precision and recall? How can AI workflows be designed to enhance user trust and explainability in high-stakes decision-making? What interaction models best support knowledge workers in integrating AI-driven insights into their workflows? The system applies domain-specific logic to critical enterprise tasks such as identifying customer pain points and drafting contracts, ensuring more coherent and traceable AI-driven decision-making. The project will evaluate the framework’s effectiveness by measuring its performance against state-of-the-art AI systems in real-world business tasks. By demonstrating improvements in reliability, usability, and user adoption, this research will lay the foundation for scalable AI-driven automation in knowledge-intensive industries. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.