AI-Powered Contract Review Assistant
AI Bot • Bot Web Portal
Design Challenge
The Web Portal Bot is an AI-powered contract review assistant designed to streamline the validation and negotiation of NDAs (Non-Disclosure Agreements). In a global and decentralized context—where contract reviews relied on slow, manual exchanges—the solution acts as a decision-support system, identifying inconsistencies and suggesting improvements to accelerate the approval process.
Methods
The project’s design goals focused on aligning cross-functional perspectives and structuring a consistent solution, following the Double Diamond approach. Through the validation of Proofs of Concept (POCs) and user needs mapping, we consolidated key insights to support product decisions and guide the creation of interfaces and interactions aligned with the legal team’s future vision.
Problem Identification
Heuristic Evaluation
Conducted with 3 designers
User Interviews
Conducted with 9 members of Dell’s legal team from different regions around the world
Heuristic Evaluation
We have listed problems, categorized by severity, to be corrected in the next phase of the project!
User Interviews
The interview analysis uncovered key opportunities to improve the NDA review workflow, particularly related to tool fragmentation, manual tasks, and limitations in automated support.
Users reported challenges with contract centralization and traceability, as requests were received through multiple channels—such as internal tools and email—impacting both workflow visibility and metrics tracking.
Another recurring issue was the presence of repetitive operational tasks, such as converting files from PDF to Word. While necessary for enabling editing and commenting, this process was performed outside the main tools, creating inefficiencies.
Additionally, the bot was perceived as disruptive to the workflow, requiring extra steps outside of existing tools. Integration with the current ecosystem—such as email and document editors—was highlighted as essential.
Personas
In this project, the team chose to use qualitative personas, based on qualitative research with smaller sample sizes, such as interviews.
The criteria used to define these personas included:
Seniority Level
Analyst, Lawyer, or Consultant
Core Activities
NDA review, contract review and negotiation, and management of the Contract Lifecycle Management (CLM) process
Location
Latin America, Europe, India, and the United States
Bot Usage Frequency
Ranging from frequent use when reviewing in-scope contracts, to no prior knowledge of the tool, basic awareness, or direct involvement in the project team
Solution Generation
We presented all research findings to the team and facilitated a collaborative workshop to generate ideas involving all project stakeholders.
6-3-5 Method
We used the 6-3-5 method, a structured ideation technique designed to engage the entire team in proposing solutions. In this approach, 6 participants each write down 3 ideas within 5 minutes. The sheet is then passed to the next person, who builds upon the existing ideas—expanding and refining them collaboratively over multiple rounds.
Benchmarking
Benchmarking results provided a broader view of existing market solutions. We focused on identifying products and approaches aligned with the ideas generated during the 6-3-5 sessions.
Solution Prioritization
User Journey
Following the research phase, we defined the user journey. Since each persona interacts with the solution with different goals and contexts, three distinct journey flows were created.
Interface
Bot Web Portal
Lab - Legal AI Bot plugin
Bot Expressions
