Allstate

Designing Decision Infrastructure at Enterprise Scale

Rather than attempting to “simplify” inherently complex systems, the work focused on stabilizing mission-critical workflows, clarifying how work moved through the organization, and reducing unnecessary variation in how similar decisions were made. A consistent throughline was reducing cognitive load in time-pressured scenarios, ensuring that people could act confidently even when information was incomplete or constrained by policy, regulation, or legacy process.

Across claims, NLP, liability analysis, and life insurance sales initiatives, I partnered closely with product management, engineering, data science, legal, and operations to support consistent, accountable decision-making at scale—designing tools that helped people move faster without collapsing nuance or displacing human judgment.

Learning at Scale: LUMA & Large-Team Collaboration

Working inside a large, distributed design organization was formative in its own right. During my time at Allstate, I became LUMA Certified, grounding my practice in shared methods for problem framing, collaboration, and decision-making. The training provided a common language that made it easier to work across disciplines and seniority levels—particularly in complex, multi-stakeholder initiatives.

(See: LUMA Institute)

Equally important was learning how to collaborate well inside a truly large system: navigating competing priorities, aligning across teams, and assuming good faith in situations where context was often incomplete. I worked alongside thoughtful leaders and peers who helped shape how I show up in collaborative environments to this day.

Working inside an XP lab environment shaped not just what I designed, but how I worked. Much of this work was done through pair design and pair programming, embedded directly with engineers and product partners. This model demanded clarity, humility, and sustained collaboration under real delivery pressure—and it significantly deepened my understanding of how design decisions translate into production systems.

I was fortunate to work alongside colleagues who modeled what principled collaboration looks like at scale:

• Katie Swindler, whose creative rigor and generosity led to an ongoing collaboration beyond Allstate, including a later book illustration project (Life and Death Design).
• Erik Pedersen and Pradeep Nayer, who consistently modeled a principled approach to collaboration grounded in assuming positive intent until proven otherwise—a mindset that proved essential in large, high-pressure systems.
• Carmelina Piedra, whose willingness to power through difficult workloads while advocating relentlessly for users and customers left a lasting impression. Her confidence in her judgment—and her encouragement to trust it—helped me learn to trust my own instincts more fully as a designer working in complex environments.

Many of the relationships formed during this period remain central to my professional and personal life today. Working this closely, under constraint and at scale, reinforced that strong systems are built not just through good design decisions, but through mutual respect, shared ownership, and sustained care for the people those systems serve.

Claims Digital Transformation

(See: Claims Digital Transformation)

I contributed to a large, cross-department effort to modernize claims intake, triage, and resolution across channels. Much of the work involved mapping deeply embedded legacy processes into more coherent end-to-end flows—without disrupting systems that were already revenue-critical and heavily regulated.

Design efforts emphasized:

• Stabilizing workflows under operational load
• Reducing friction in high-volume, time-sensitive tasks
• Clarifying handoffs and decision pathways across teams
• Improving usability where adjusters faced the highest cognitive burden

The goal was not reinvention, but making existing work more reliable, consistent, and inspectable—particularly in edge cases and exception handling.

Claims NLP & Decision Support

(See: Allstate Claims NLP)

As Senior Designer on claims NLP initiatives, I worked on tools that applied statistical language processing to claims documentation and analysis—well before today’s generative AI framing became common.

Rather than treating NLP as authoritative, we positioned it as assistive decision support. Design work focused on surfacing uncertainty, preventing over-reliance, and ensuring human judgment remained central.

Liability Analysis Tool

(See: Allstate Liability Analysis Tool)

I also served as Product Designer on a liability analysis tool supporting fault determination—one of the most judgment-heavy aspects of claims handling. The work emphasized explicit reasoning, reduced variance, and clear accountability between human and system.

Life Sales Central

In parallel, I contributed to Life Sales Central, a life insurance application used by both internal sales teams and customers.

Field research revealed that agents frequently turned their screens toward customers to explain policy differences—using the interface itself as a shared reasoning surface. This insight shaped design decisions around clarity, cognitive load, and trust in live, high-stakes conversations.

Impact

The work contributed to measurable operational improvements, including multi-million-dollar annual savings and double-digit reductions in handling time per claim, as reported by internal program metrics. These outcomes were achieved not through full automation, but by improving the reliability, usability, and consistency of tools used daily by claims and sales teams—particularly in judgment-heavy workflows.

Reflection

Designing inside a 90-year-old enterprise clarified that progress is rarely blocked by technology alone. More often, it is constrained by trust, accountability, and institutional memory. My role at Allstate was less about novelty and more about stewardship—learning how to intervene thoughtfully, collaborate generously, and improve outcomes without destabilizing systems people rely on.