Microsoft Contracts

1 million contracts, one modern platform

92

Microsoft’s platform for managing over 1 million contracts had become outdated and hard to navigate, forcing teams to spend significant manual effort just to find what they needed. I redesigned the entire experience end to end and introduced an AI agent to make finding and managing contracts fast and intuitive, now serving 30,000 visitors a month.

Role
Lead Product Designer
Years
2026
Scope
Strategy, UI/UX Design, Research
Microsoft Contracts — redesigned dashboard with the agent ask bar and portfolio stats
Microsoft Contracts — AI Contracts Agent conversation

Challenge

Microsoft Contracts was outdated and hard to use, yet legal teams relied on it daily to manage over 1 million non-revenue contracts. With no dedicated researchers, I had to learn complex legal workflows before redesigning it, understanding how teams actually worked before changing anything. The challenge was to modernize a deeply embedded, high-stakes system without disrupting its users.

Impact

0 Million+

Contracts Managed

0+

Monthly Visitors

0% Faster

Task Completion

0+ Hours

Estimated Annual Savings

Collaboration

As lead product designer, I collaborated directly with business stakeholders, PMs, and engineers from vision through execution, keeping design aligned with business goals and technical feasibility at every step. I also owned the research myself, ensuring the redesign was grounded in real user workflows rather than assumptions.

Designing for an Existing Userbase

Microsoft Contracts manages every non-revenue contract at Microsoft, and two groups do that work. Paralegals handle most of the daily tasks, while attorneys manage teams of paralegals, and the platform is designed around each of their workflows.

User Type 1

Paralegals

Paralegals carry the daily volume, drafting agreements, chasing renewals, and pulling terms for whoever asks. Their work is fast, repetitive, and mostly about finding the right contract quickly.

User Type 2

Supervising Attorneys

The attorneys managing those paralegals rarely open a contract themselves. They need the state of the whole book, what’s due, what’s stuck, where the risk sits, and how the work is spread across their team.

Talking to Users and Stakeholders

I interviewed users to understand where the previous platform failed them, and met with business stakeholders to collect all necessary requirements. I took both perspectives into account with every design decision as I pushed the project forward.

Questions for Users / Business

“Can you walk me through how this platform is used today?”

“What is the current platform missing that you need most?”

“What specifically would you want an AI experience to help you with?”

“What current functionality would you not want us to disrupt?”

“If we could only fix one area of the platform, what would it be?”

“Where is time being lost across the current platform?”

“Where does the existing platform slow you down the most?”

User Pain Points

From several interviews with paralegals and attorneys

Business Requirements

From planning meetings with CELA business stakeholders

Research synthesis — user pain points from paralegal and attorney interviews beside business requirements from CELA stakeholder meetings, with the same missing-AI finding starred on both

Auditing the Existing Experience

After the business stakeholder conversations and user interviews, I audited the existing experience, documenting where it failed users and what legacy experiences needed to be carried forward.

Audit of the old Legal Contracting Experience — the legacy dashboard annotated with notes marking failures and the legacy pieces worth keeping

Mapping the New Experience

I mapped every feature of the new experience from three inputs, user pain points, business stakeholder requirements, and the old platform itself. The PM and I used this map to align on strategy and direction before design began.

New experience map — the platform's features organized by nav item, traced from pain points, stakeholder requirements, and the old experience

Sketches & Wireframes

I started with rough sketches to explore different structures, then progressed my preferred directions into low-fidelity wireframes. Along the way I reviewed the work with business stakeholders, validating direction while changes were easy and quick to make. By the time I moved into high-fidelity design, the direction and structure were already agreed on.

Initial dashboard sketches — versions 1 and 2, exploring AI entry points, key actions, and contract views
Dashboard sketch version 3 — actions, stats, and contracts preview, with the AI chat as the first element on the screen
Low-fidelity dashboard wireframe — greeting, AI entry, key actions, contract statuses, and the contracts preview list

Initial Sketches

Dashboard Wireframe

Key Design Decisions Based on Research

These are some of the key decisions I made on the new dashboard experience, directly informed by talking to users, as well as the business requirements and the audit of the existing legacy platform.

The redesigned contracts dashboard — AI entry point up top, stats, and the contract list

AI Entry Point

First and most prominent element on the dashboard

Encourages users to engage with it right away

Legacy Experience Entry Points

All critical legacy experiences still have entry points

Clear but not overly emphasized

Statistical Overview

Quick big-picture overview of their contracts

Ability to change view and see team members' workload

Important Contracts

Surface relevant contracts right on the dashboard

Metadata overview, AI summary, and the ability to view

Before — the old Legal Contracting Experience portal
After — the redesigned contracts dashboard

Before

After

Key Features

Emphasis on AI Experiences

The agent was the single most important goal for the business and the biggest addition for users, so it got more time and detail work than anything else on the platform. Most of that went into the output itself, how answers read, what they cite, and how they hold up when a paralegal acts on them. Done right, it saves hours of manual work.

Custom Components for the Contracts Agent

I designed five custom adaptive cards specifically for the Contracts agent, each built around a recurring moment in the contracts workflow that users encounter.

Contract cards — renewal summaries with supplier, amount, and owner
Contract reference card — a cited clause with the operative language highlighted
Timeline card — a contract's renewal path with milestones and the gating step
Person card — contract ownership and workload at a glance
Comparison card — two contracts side by side

Building the Design System

I put together a design system for Contracts that sits within Microsoft's global Fluent 2 system. It prevents visual drift, keeps the platform easy for other designers to pickup and maintain, and makes handoff simpler.

Dashboard Components sheet — stat, action, and template cards, the dashboard composer, form fields, progress steps, contract summary and list, buttons, filters, and tabs
AI Agent Components sheet — the agent composer, action and suggested-prompt chips, contract, timeline, comparison table, person, and clause excerpt cards

AI Making an Impact

This is the technical workflow behind the project. AI helped collapse the gap between design and code. I moved from Figma designs to a working demo fast enough to get business sign-off on the real thing, then handed the frontend UI straight to engineering.

The AI workflow behind Contracts — Figma designs through Claude Code to GitHub, deployed to an Azure sandbox for stakeholder demos and frontend handoff to the SWE team

Colleague Feedback

Feedback from leaders and partners I worked with on the project, shared during performance reviews.

Charles stepped in without hesitation and delivered. He brought a fresh and thoughtful perspective on every request, proposed multiple well-considered design options, and turned things around at an incredible pace.

Microsoft

Senior Product Manager

Microsoft

A standout strength of Charles is his ability to deeply understand and balance both PM and SWE needs. His collaborative mindset, responsiveness, and design leadership made a significant difference.

Microsoft

Principal SWE Manager

Microsoft