Background of application
Lawyers waste significant time and resources on analyzing and reviewing large volumes of contracts
Identify Pain Points
Solution: AI Agent for Legal Contract
✅ Save 50% drafting time, Reduce errors from manual citation checks✅ Decreased legal documents review time by 34% through a streamlined AI driven process, saving 1,200 hours monthly.
✅ Boosted accuracy of reviewing documents by 67% , which helped in generating additional revenue.
Design Process
01 / Empathize – Understand users
user research
Through user research I tried to understand how legal professionals handle contract drafting, reviewing, and negotiation—and evaluate opportunities for an AI assistant to improve efficiency, accuracy, and client outcomes.
✔ 5 rounds of usability testing with 87% task success rate
✔ 12 in-depth user interviews -experience 2–20+ years (avg. 9 years) across 3 user segments
✔ 3 prototype iterations with 25 participants each
✔ 2,000 h session recordings analyzed
KEY FINDINGS : USER RESEARCH
Drafting Time Lawyers spend 38–45% of their week drafting and revising contract
Review Workload 70% said contract review is highly repetitive and time-consuming.
Negotiation Challenges 61% struggle to track clause changes or non-standard edits.
Error Risk 43% have missed a risky clause due to time pressure.
Expected ROI Predicted 20–30% time saved in drafting and 25–40% in review.
OPPORTUNITY INSIGHTS : USER RESEARCH
High willingness to try AI (78%)
Time savings (30–40%)
Strong market gap
COMPETITIVE ANALYSIS
- AI drafting & summarization
- Law firm integration
- Limited explainability
- Not customizable for smaller firms
- Fast clause drafting
- MS Word integration
- Narrow scope
- Limited negotiation support
- Advanced machine learning
- Due diligence & analytics
- Expensive & complex onboarding
- Designed for large enterprises
- End-to-end workflow: Draft, Review, Negotiate
- Explainability & transparency
- Customizable clause library
- Mid-size firms and corporate legal teams
- Negotiation analytics
02 / Define – User’s Needs and Problems
From the Empathize/Discover stage, I gathered and clustered findings from Lawyers, paralegals, procurement managers, compliance officers, startup founders, etc which in turn helped me to write a clear Problem Statement or How Might We question
how might we (hmw questions)
💡 How might we make AI review workflows align with existing contract management tools?
💡 How might we help users quickly understand a contract’s key terms and risks?
💡 How might we ensure AI-generated insights are legally trustworthy and explainable?
🎯 Objectives and Success Metrics
Objectives Reduce time to review contracts
Success Metric 50% reduction in average review time
Objectives Improve clause consistency
Success Metric 80% accuracy in clause detection
Objectives Build trust in AI output
Success Metric >90% of users rate clause summaries as “reliable”
PERSONA
Based on my research, I found some trends among my participants. One group of participants consisted of corporate lawyers.
Meet Amit Jain, corporate legal head in a reputed mid sized company
👤 Amit Jain
Occupation
In-house counsel at a mid-sized tech company
Demographics
- 34 years old;
- Lives in Delhi;
- IIM from Madras
- Work Environment: Hybrid (office + remote)
- Tech Comfort High (familiar with SaaS tools, AI assistants, and collaboration platforms)
- Has an upper-middle-income level
Motivations 🏅
────୨ৎ────High-quality outputs efficiently Boost productivity easily. Appear professional and reliable to clients
Goals and Needs 🎯
────୨ৎ────
Reduce time spent on contract review.
Ensure compliance with internal legal policies.
Support non-legal teams (sales, procurement) in drafting contracts.
Pain Points🤔
────୨ৎ────
- Repetitive clause reviews across NDAs, MSAs, and SLAs.
- Frustrated by errors and inconsistencies
- Delays due to manual contract redlining..
- difficult to scale personalized documents
🤖 How the AI Agent Helps
Automated clause analysis and risk scoring
Compliance checking against internal playbook
Redline suggestions with legal rationale.
04 / Prototype
05 / User Testing
I have designed user testing to evaluate how effectively and efficiently real users can perform their tasks using the AI agent compared to manual review.
✓ A/B Testing, Remote Testing, Quantitative Evaluation
✓ 4 variants tested across 500 users
✓ 95% confidence level in results
✓ 2-week test period per variant
✓ 14 key metrics tracked
- 5-10: Lawyers, paralegals.
- 5 – 10: Managers who deal with contracts
- Optionally 5 non-legal testers
- Identify high-risk clauses.
- Summarize the contract in plain language
- Suggest improvements and rewrite
- Check for missing clauses
- Accuracy
- Precision / Recall
- Task Completion Time
- User Satisfaction Score (Likert scale)
- AI reduces review time by 35%.
- AI precision = 92%
- Users rated AI usefulness 4.5/5 on average
Leave a Reply