Product
High Q, STWB, TxM, Pendo
Services
UI/UX, AI Research, Facilitator, Information Architecture, Design System
Responsibility
Product Lead
As a seasoned Senior UX Designer and Product Lead, I've had the privilege of working on groundbreaking projects like Reuters' HighQ AI and Document Intelligence. During my tenure with Reuters, I facilitated the end-to-end development of an AI-driven contract analysis tool tailored for legal professionals. This role demanded a deep understanding of the legal landscape and an ability to harness technology to streamline complex workflows and empowering lawyers to identify key risks and obligations more efficiently. Moreover, I led the creation of intuitive user interfaces, leveraging my UX expertise to enhance user experiences and ensure seamless communication of complex legal concepts. My contributions to HighQ, particularly in areas such as document automation, workflow optimisation, and secure data handling, significantly improved operational efficiency, reduced risk, and ultimately elevated client service standards. An 8 hour contract review process was reduced to 2 hours, time savings of at least 72%.
Challenge
Legal teams spend countless hours manually reviewing lengthy contracts. The process is time-consuming, prone to human error, and delays decision-making.
Process
MVP Development Process
1. Define MVP Scope
Focused on key clauses and risk flagging
Targeted for legal teams handling NDAs and MSAs
2. Data Collection & Labeling
Gathered anonymized contract samples
Collaborated with legal SMEs to annotate critical clauses
Tools: Google Drive, Labelbox
3. Data Preprocessing
Extracted text from contracts (PDFs, DOCs)
Cleaned and standardized data for NLP processing
Tools: Python (Tika, PyPDF2), Tesseract OCR
4. AI & NLP Pipeline
Built a clause extraction engine using LegalBERT
Implemented Named Entity Recognition (NER) for legal entities
Tools: Hugging Face Transformers, SpaCy
5. Model Training & Evaluation
Fine-tuned pre-trained models on contract data
Validated with legal SMEs for accuracy
Tools: PyTorch, Scikit-learn
6. Backend API Development
Created REST APIs to analyze contracts and return extracted insights
Tools: FastAPI, Python
7. Frontend Dashboard
Designed a simple UI for contract upload and analysis results
Displayed flagged clauses and risks for quick review
Tools: React, Material UI
8. Testing & Deployment
Conducted end-to-end user testing with legal teams
Deployed MVP on AWS with secure access
Tools: AWS, GitHub Actions for CI/CD
Tools & Technologies
NLP/ML: LegalBERT, ContractNLI, Hugging Face, PyTorch
Data Processing: Python, Tika, PyPDF2, Tesseract OCR
Annotation Tools: Labelbox
Backend: FastAPI
Frontend: React, Material UI
Cloud: AWS
CI/CD: GitHub Actions
We developed an AI Contract Analysis MVP to a full Customer Launch product that could:
Automatically extract key clauses (Termination, Liability, Jurisdiction, etc.)
Flag potential risks
Provide an intuitive dashboard for quick review
Impact
Delivered a working MVP in 12 months
Reduced manual contract review time by an estimated 72%
Secured buy-in from legal stakeholders for further product development
Laid the foundation for Thomson Reuters’ future AI-driven legal tech solutions
