Zac Halbert

Zac Halbert

Product Design Portfolio

Metromile

Head of Design, Enterprise (2018 – 2021)

I was the founding designer and Head of Design for a new organization within Metromile that aimed to harness Metromile's expertise in data science to offer solutions to multinational insurance organizations. My work directly led to 7 figures in revenue within the first two years, and saved our customers a similar amount.

Some details have been obfuscated for NDA reasons.

Metromile Enterprise
I Led a Design Sprint to craft a first notice of loss flow with a large group of executives in Tokyo

My Role

I joined Metromile to help get its internal insurtech off the ground. I hired a team of two designers, and worked closely with engineering and data science to offer three complete products: Report, a first notice of loss product, Detect, an AI-powered fraud detection solution, and Streamline, an insurance-specific workflow automation engine. I also led the branding and visual design efforts.

This is a handful of the skills I regularly practice in this role:

Team Management

Hired and managed a scrappy design team of two, both who have gone on to lead design teams.

UX and UI

Designed all user flows, app UI, and design system in both Figma and code.

Product Vision & Strategy

Collaborated with founders to set overall product vision and vetted vendor selection.

User Research

Conducted user interviews and contextual inquiry to understand user needs, goals, and product usability.

Collaborative Design

Led multiple design sprints onsite with international companies to increase speed and quality of feedback.

Rapid Prototyping

Rapid prototyping in Figma to quickly get feedback from users and technical experts.

Brand Design

Designed the brand, logo, voice and all print collateral.

Frontend Development

Regularly shipped code to production using CSS and React.

The Challenge

Designing for multinational insurance companies proved to be a rewarding challenge. Many of these companies had eye-watering amounts of revenue, but were culturally very conservative. Some of the youngest insurance companies have been around for at least a generation, and are accustomed to upgrading their technology on a cycle measured in decades rather than years. Approaching these types of companies with cutting-edge AI powered solutions proved to be quite difficult.


My Approach

One major benefit we had was the ability to point to the success of Metromile's consumer product, which was powered by the underlying technologies that we were offering. I used a variety of collaborative design techniques like Google Ventures Design Sprints, and built on top of existing models (Business Process Modeling Notation, or BPMN) to increase the trust in our offerings.

The design process was a typical human-centered, rapid prototyping based method. To get feedback, I often needed to travel to client locations such as Tokyo, Japan, to get in-depth feedback on solutions.

Products

What follows are a few of the major products I worked on. Many of the details cannot be shared for NDA-related reasons, but I can provide a high-level overview of each.

Product

Streamline

Streamline was a workflow automation product that allowed insurtech companies — which operate on razor-thin margins — to automate repetitive tasks for their claims adjusters and internal processes.

This product enabled double digit efficiency gains and a large reduction in errors by automating repetitive, error-prone, and mundane tasks. We used the human-in-the-loop methodology to keep people a part of the process, allowing them to observe and direct the automations at will.

Metromile Enterprise Streamline UI
Product

Detect

A double digit percentage of all claims are fraudulent, and insurance companies are compelled by law to investigate and report these crimes. This is an enormously difficult and costly process. Metromile Detect used SHAP analysis and proprietary machine-learning models to offer investigators insight into which claims might be fraudulent, with clear steps to push each case forward.

Metromile Enterprise Detect UI
Product

Report

When an insurance customer has an accident, the first step is to report it. This is typically called first notice of loss (FNOL). Report was a dynamic FNOL product that allowed users to craft a dynamic flow based on the specific type of incident.

This dynamic form-builder allowed one major customer to drastically decrease the time it took for their customers to report a loss, since customers were only asked questions that were relevant to their situation. This led to an increase in completions and better data for claims adjusters.

Metromile Enterprise Report Designer UIMetromile Enterprise FNOL UI

Branding

As the founding designer, I was responsible for crafting the brand and logo for the enterprise organization. I worked cross-functionally with stakeholders to understand the vision for the organization as a part of this process. The goal was to make the brand feel like a subsidiary of Metromile Consumer, with a heavier emphasis placed on elements that would work better for enterprise customers.

Metromile Enterprise Logo DesignMetromile Enterprise Marketing Website

Reflections

This role was challenging, but incredibly fun. Some of my most rewarding memories were running design sprints in Tokyo with 25 executives and a couple translators, then going out to celebrate our partnership afterwards.

I also ended up learning a great deal about business process automation, and became trained in BPMN. This is somewhat similar to designing user and task flows, but with much more precision and a richer set of interactions. BPMN has become a part of my usual product design process when I need to express a complex series of interactions for many actors. In many ways, BPMN is a slightly more complicated version of many of the deliverables found in the Service Design discipline.

It was rewarding to work on so many flow-based products. The challenge was figuring out how to design a product that kept humans at the center while allowing them to automate the stuff they aren't interested in doing. In addition to learning a lot about AI and automation, I learned a lot about the enterprise sales process, especially when working internationally.

Next Project:

Samahope