International Trading Company
Integrated portfolio management and AI-powered contract intelligence for an international trading company
Summary
Transforming how a trading organization manages its portfolio and processes contracts through unified digital tooling and intelligent automation.
The company managed one of the world's largest energy product portfolios through fragmented, legacy-based workflows. Teams across multiple locations relied on disconnected tools and manual processes. We partnered with internal teams to design, build, and deliver both a unified portfolio platform and an intelligent contract extraction system.
Services
Industry
The Challenge
Disjointed systems, manual workflows, and contracts too complex for humans to process efficiently.
The company managed one of the world's largest energy product portfolios through fragmented, legacy-based workflows. Teams across multiple locations relied on disconnected tools and manual processes, limiting visibility, slowing decision-making, and making it difficult to unlock new commercial opportunities. At the same time, contracts often exceeded 200 pages and required significant hours of manual data entry.
Problem 01
Fragmented portfolio management
- ▪Legacy tools spread across multiple locations
- ▪No unified view of the global portfolio
- ▪Computationally heavy models with long runtimes
- ▪Disparate data sources and siloed teams
Problem 02
Manual contract processing
- ▪200+ page contracts processed by hand
- ▪10+ hours of manual data entry per contract
- ▪Highly variable document structures
- ▪Error-prone extraction of critical terms
The Approach
We collaborated closely with internal teams to deliver two interconnected workstreams: a unified portfolio management platform and an AI-powered contract intelligence system, both designed to scale with global trading operations.
Workstream 1: Portfolio Platform
01
Backend consolidation
Built scalable infrastructure for computationally heavy models, unifying multiple data sources and legacy systems into a single, performant platform.
02
Consistent UX/UI across five business areas
Designed a consistent interface for different roles and locations, ensuring the platform worked intuitively for teams across the organization.
03
Bespoke scheduling solution
Co-designed and iterated on a scheduling solution with end users, optimizing how trading teams plan and execute across the portfolio.
Workstream 2: Contract Intelligence
01
AI-powered data extraction
Developed intelligent extraction using RAG and multi-agent architectures to automatically process uploaded contracts and extract structured information with high accuracy.
02
Human-in-the-loop validation
Built a purpose-built interface where users can accept or reject extracted fields individually, mitigating risks of AI hallucination while preserving efficiency and control.
03
Contract querying and summarization
Integrated an AI-powered chatbot for natural language contract querying and automated summarization, enabling teams to surface key terms without manually reading hundreds of pages.
Impact
A unified platform that empowers trading teams globally, and an AI system that turns weeks of manual work into minutes.
11 roles across 3 locations
Brought together on a single, cohesive platform, replacing fragmented legacy workflows.
50% reduction in model runtime
Through a scalable backend architecture optimized for computationally intensive operations.
200+ pages processed per contract
Automatically extracted, structured, and validated from complex, variable-format commercial agreements.
10+ hours saved per contract
Of manual data entry eliminated, while improving accuracy and creating a searchable contract repository.
Methods
A collaborative design and engineering approach, combining service design with advanced AI architecture and human-centered validation.
Journey and service mapping
RAG pipelines
Multi-agent architecture
Prompt engineering
UX/UI design
Co-design workshops
Human-in-the-loop design
Usability testing