About
A US-based company founded by former Amazon and Microsoft engineers was developing a SaaS platform for construction and legal teams to streamline contract analysis. With the LLM-powered solution developed, they achieved 70% faster processing and 90% higher accuracy across all document types.
Challenge, approach, and impact
LLM-Powered System for Contract Analysis
Development of an LLM-powered system to automate contract analysis workflows for construction and legal teams, achieving 70% faster processing and 90% higher accuracy.
Challenges Faced by the Client
Addressing slow, inconsistent document reviews during the development of a SaaS platform for contract analysis, leading to delays in processing client requests and quality issues.
Sub-Challenges
1. Model Training and Quality Assurance: Multiple training iterations required for accurate analysis. 2. Speed Optimization: Ensuring fast results without compromising quality. 3. Document Type Differentiation: Correctly identifying and analyzing various contract types.
Solution Implemented
Implementation of an LLM-powered engine with decision-tree logic for automated and structured contract analysis, providing tailored prompts for each contract type and delivering structured insights within seconds.
Sub-Solutions
1. Model Training and Quality: Conducting fine-tuning iterations for consistent results. 2. Speed Optimization: Refining prompt flow and decision tree architecture. 3. Document Type Differentiation: Guiding the LLM with a custom decision tree for contextual accuracy.
Results Achieved
Achieving 70% faster document analysis and 90% greater accuracy through LLM-driven automation, leading to improved speed, consistency, and scalability of document processing within the client's SaaS platform.
How we built
Testimonials
Anonymous
Dataforest
“Dataforest helped us integrate an LLM-powered document analysis engine into our SaaS platform, significantly improving both speed and accuracy. Their team quickly understood our domain challenges, delivered a scalable solution, and optimized performance for real-world usage. As a result, we reduced document processing time by 70% and achieved consistent, high-quality analysis across all contract types.“
Team structure
Client team
Michael Johnson
Project stakeholder
Project stakeholder
Sarah Clark
Daily point of contact
Daily point of contact
The client stakeholders were working closely with the team at Dataforest
Agency team
3 x Machine Learning Engineer
Production
3 x Data Engineer
Production
2 x Business Analytics
Production
2 x Product Manager
Production
