Conceptual redesign for Daisy’s risk management platform


Over the summer of 2021, I had the opportunity to join Daisy Intelligence Inc. as a product design intern to work on the Insurance and Retail platforms, with a focus on Insurance. Regardless, I got to learn about issues that investigators and grocery managers face when it comes to navigating these platforms.

Based in Toronto, Canada, Daisy is an AI software-as-a-service company that analyzes massive amounts of transaction and operational data to help retailers and insurance companies make better operational decisions to improve their business.



Summary


Timeline
8 weeks

Responsibilities
Interaction design, visual design, user research, UI audit
Type
Internship (Remote)

Role
1 product designer with 2 engineers and 2 product managers


Introduction to the Alerts and Triage Sheets


Over the span of 8 weeks, I worked with the Daisy Insurance team to design features that made it easier for investigators to manage their clients’ information on their platforms, with a focus on Alerts and Triage Sheets.

My Role
I was the main designer of this project, in which I collaborated with the PMs and engineers to establish design requirements, build wireframes, and prototypes in order to conduct design reviews to gather feedback.

What are the Alerts and Triage Sheets?
The Alerts and Triage sheets are features of the Daisy Insurance Platform that are important for investigators to know how to prioritize cases:

  • Alerts Sheets allow SIU (Special Investigations Unit) teams to quickly understand the key reasons for suspicion of each subject scored with the Daisy Suspicion Index.
  • Triage are generated by the Daisy system to highlight the most suspicious actors using the Daisy Suspicion Index (DSI) which is a risk score out of 1000 that indicates the degree of suspicion of claims, insured members, providers, and networks.

Problem

Adjusting to a more modern & intuitive user experience 


After conducting a user interface audit and addressing client concerns, there were three main problems that needed improvement in the insurance platform:

Minimal regard for designing for users
The team recognized that the current webs platforms have been built out of necessity by engineers, with minimal regard for design and user adoption. Without the close and frequent collaboration between designers and engineers:

  • Designers: risk not having their research implemented into the final designs or end up designing for the sake of looks rather than functionality
  • Engineers: may have to rewrite a feature because it was poorly understood when it was first created

Information Overload
Being exposed to an abundance of information at once can negatively impact employees’ performance and wellbeing at work. 

Outdated design interfaces
A lack of platform consistency with design elements may be a source of friction when users expect the platform to act one way, but have it act a different way. Additionally, the overall UI isn’t the most welcoming or modern, which hinders the ability of investigators to scan and comprehend the information that is presented in front of them.

Design Goals

Ensuring efficiency for investigators


With the aspiration for Daisy to become the insurance software platform of choice for fraud prevention and claims handling, we believe that we can achieve this through a simple, intuitive, and modern user experience.

How might we redesign a product that increases the efficiency of insurance investigators as they manage large sets of data?



Our mandate was to create a simple, intuitive, and modern user experience that allows SIU and claims handling teams to:
  • Execute their work in less time
  • Adding interpretability to the outputs of the AI system 

Competitive Analysis

Understanding other Insurance Software Platforms


With the rise of other insurance tools, I wanted to get insight into the landscape of insurance platforms and how they help investigators to prevent fraud and handle claims. I explored further into other new and established insurance tools currently available.

To get inspiration outside of insurance, I also looked at other platforms that allowed users to manage large sets of data such as Guidewire ClaimCenter and Applied Epic.

Other insurance platforms

Guidewire ClaimCenter
                                     Guidewire ClaimCenter


Takeaways
  • Tables are very efficient for laying out pieces of data while giving the ability to change the view to the outcomes (case management, detailed reports, dashboards)
  • Investigators are able to connect multiple subjects to a single case or open multiple cases against a single subject over time by using specific filters.
  • Dashboards provide managers and executives visibility into investigative team throughput, fraud savings and usage of the software platform.

Ideation

Ideating within and beyond the design system


Using the takeaways from the competitive analysis, I created a few low-fidelity wireframes in an attempt to apply those concepts to the Insurance Platform. 

Explorations

Finding a balance between modern and efficient


Issue 1: Lack of Hierarchy in Navigation Bars
  • One of the main issues was that the primary navigation bar was on the right side and the secondary bar was on the left; at least in Canada and the USA, the text is usually read from left to right. Therefore, the hierarchy of navigation was not intuitive to new users.


Issue 2: Overabundance of information in tables 
  • The second exploration was that to investigators, especially new ones, was that there was often an overwhelming amount of information that was presented to them in the tables at once.  



Issue 3: Lack of Content Separation + Cluttered UI
  • The last exploration was into the outdated design interfaces; there was a lack of content separation within the current card designs and the platform was often overwhelmed with different design elements. This exploration was chosen because it addresses the two design goals most efficiently. 


Iterations

Refining UI and Information Architecture


After settling on a direction, my next steps were to work with the product management and engineering teams to conduct design critiques and flush out mockups and components. Through close collaboration with these teams, I could get a better understanding of which designs were the most feasible and take any constraints into account.

Restructured Information Architecture (Triage)
To help the platform become more intuitive for new and returning users, I updated the information architecture of the navigation bars.

Pros
  • More intuitive navigation of where information can be found and accessed (creates information hierarchy)
  • More room for the top-level links that have long names
  • Helps to differentiate which links are primary and secondary
  • Labelling of content in an effective and sustainable way
Cons
  • To keep the interface as modern as possible, many of the pages had to be combined into fewer page titles, meaning that users would have to relearn where the information was stored


Prioritized and Hidden Table Data (Alert)
In an attempt to reduce the information overload for investigators, I proposed the idea of reducing the main table into a condensed version with a pop-out card a pop-out card is included if more details are needed. Additionally, the majority of the table filters are hidden inside of a drop-down tab.

Pros 
  • The condensed table helps investigators to quickly scan for necessary information
  • The overall layout of the platform is more modern and less cluttered
  • Reduces information overload when looking at the tables because only the most essential information is shown
Cons
  • The separate filter menu seems hidden and it may decrease the efficiency of investigators if they have to change the filters every time



Card Design Implementation (Triage + Alert)
To increase the consistency of the UI design elements and modernity of the current platform, I sorted the information into cards.

Pros
  • Allows investigators to have their information sorted in an organized manner
  • Cards make the content on the page easily scannable
  • Easier to create a single aesthetic across multiple devices (as one of the future aspirations of Daisy is to move to a more iPad/mobile experience for investigators 
Cons
  • The transition to revamp the entire web platform is costly
  • Lack of hierarchy in the information on the page may confuse users that are looking for a path to complete their task rather than wander randomly


Final Designs

Allowing investigators to efficiently scan and take action on cases


Wrapping up the final iterations, I developed some conceptual sketches/prototypes in Figma for designers, engineers, and product managers to test interaction patterns before refining the small details for the hand-off.

Triage Sheet                                 Triage Sheet

Reflection

Optimizing a remote design internship


This was my first design internship and I was able to learn a lot about the different product offerings of Daisy, how the team works together, and the kinds of challenges that investigators face in all stages of the business; I also learned a lot about both the insurance and retail industries through this experience.

I’m excited that I had the experience to:

  • Explore more blue-sky designs that push engineering constraints to accommodate client/investigator requests for a more seamless and user-focused journey
  • Share with the team about new design tools like Figma. The prototypes helped the team communicate ideas more clearly to other non-technical company members
  • Share my internship findings and designs with a board of key members from Daisy