Driver Risk Score
Samba Safety + Crafted
Overview
Samba Safety
Driver risk management platform, serving companies with professional drivers: gig platforms, long-haul trucking, corporate fleets.
Top 25 customers represent over 1 million drivers.
Driver Risk Score
Samba's main users are Driver Safety Managers who are responsible for driver risk and compliance at their companies. The Driver Risk Score gives them one trusted number to act on for every driver, built from a proprietary algorithm inside Samba.
How It Works
A Motor Vehicle Record (MVR) is the license and violation history of a driver. It has long been the industry's main safety metric, and it shapes insurance premiums, employment eligibility and overall trustworthiness of a driver. The Driver Risk Score combines MVR with on-the-job incident data. Safety Managers use it to recognize safe drivers, assign training or flag a driver before rising risk turns into a costly incident.
My Role
Lead Designer. Owned design and research end-to-end, helped define hiring criteria and mentored Samba's first in-house designer, and led stakeholder alignment across product, sales and leadership.
Current State
Before this project, driver safety data was scattered and difficult to act on, forcing Safety Managers to make high-stakes decisions on insufficient information:
-
Disconnected data: Official violation history (MVR) lived in Samba, but incident data (claims) was in spreadsheets, email threads and other tools
-
Incomplete picture: MVR doesn't capture every incident a driver may have on the job, and claims data wasn't quantified in a way Safety Managers could confidently act on
Bottom line: Unsafe drivers went unnoticed, and every incident costs the company money.
Problem to Solve
How can Samba provide Safety Managers with one complete picture of driver risk, so they can reward the drivers keeping people safe and stop the ones who aren't?
Project Goals
Goal #1
Move Incident Claims into Samba without disrupting how teams already work.
Challenge + Response
-
Customers wanted claims data in Samba, but years of using an existing process made adoption a real concern.
-
Designed a bulk upload tool with a structured CSV template for fast, low-friction migration.
Goal #2
Help Safety Managers trust and adopt an entirely new way to understand risk.
Challenge + Response
-
Limited user access meant every touchpoint had to count, and the score introduced a concept new to the market.
-
Ran a targeted survey to focus research time, then built Pendo onboarding and tooltips to introduce the score in context.
Goal #3
Align with business teams to build Claims for today, while also designing towards the future.
Challenge + Response
-
Sales had already promised capabilities that didn't exist yet, creating pressure for this project's timeline.
-
Partnered with sales and support to balance previous commitments with what users needed most, and not compromise the next project.
Research & Discovery
User Personas
Driver Safety Manager
Acts on driver risk: assigns training, flags concerns or takes a driver off the road. Making high-stakes decisions every day that protect drivers and the public.
Driver
The professional driver behind the wheel. Doesn't interact directly with Samba, but every policy, score and decision exists to keep them, and the public, safe.
SME & Stakeholder Interviews
Subject matter expert (SME) interviews helped us ramp up fast on a specialized, regulated industry. Stakeholder interviews ran in parallel, aligning on goals and success metrics that shaped the project's direction and research plan.
Survey
Designed and sent a survey to 1,600 VP, EVP and Director-level contacts across Samba's customer base. This provided quantitative signal and leadership sentiment at scale to de-risk assumptions and validate direction. From there, interviews focused on Safety Managers and their day-to-day work.
The survey was built to test the assumptions above
Key Findings
75%
were already tracking claims data outside Samba
80%
were analyzing or wanted to analyze claims alongside MVR
70%
said combining the data would be extremely valuable
61%
wanted company-wide safety metrics visible inside Samba
User Interviews
Direct conversations with Safety Managers about their daily workflows and how they evaluate risk today:
-
How do they currently track and act on incident claims outside Samba?
-
What's the source of truth they rely on to evaluate a driver's risk today? What is missing?
-
How do they define and update safety policy?
-
What would make combining MVR and claims data valuable enough to justify a workflow change?
This early research was synthesized and weighed against engineering effort to decide where to focus first. The user insights carried forward into design decisions throughout the ideate and iterate phases.
Ideate & Design
Design Studio
A "How Might We?" workshop brought the CEO, CTO, EVP of Product, EVP of Sales, and representatives from the Marketing and Data teams into the room early, giving stakeholders a rare chance to share and shape ideas they'd been developing individually. The session generated a wide range of concepts. From there, I synthesized the strongest ideas into a direction worth testing with users.
From Sketch to Screen
Low-fidelity sketches, created together with engineers, kept the team anchored on feasibility and flow before spending time on any visual polish. From there, I moved straight into high-fidelity Figma designs by extending Samba's existing component library and building new elements only where Claims and the Driver Risk Score demanded them.
Claims
Claims sketches
Driver Profile sketch
Adding scattered incident history into Samba raised usability concerns that needed to be addressed through the design process and into the launch.
Driver claims
Migration
Friction
Manually entering years of records wasn't practical at the scale most customers needed.
Fix
A bulk upload tool with a structured CSV template let companies migrate existing history in one pass.
Adoption
Friction
Real change is difficult: breaking habits, adjusting internal process, trusting a new system with critical decisions.
Fix
Released product with a communication plan to ease the transition from customers' existing tools into Samba.
Automation
Friction
Scoring every claim individually would defeat the purpose of a bulk upload.
Fix
Default rules by claim type were applied automatically at upload. This consistency was crucial for building trust in the new score.
Reporting
Friction
Customers needed a fast way to see claims data in aggregate: totals, costs and trends over time, not just individual records.
Fix
A data view that tracked overall claims activity, a first step toward the fleet-level reporting we would later build on.
Driver Profile Improvements
Before
Showed only a driver's MVR score, color-coded by risk level, but an incomplete picture Safety Managers had to work around.
After
One view combining MVR and Claims scores into a composite Risk Score, where a higher score means higher risk, plus a full Activity History showing exactly where each data point came from.
Onboarding
Pendo introduced the new score right where Safety Managers first encountered it.
Before
After
Note: Some mocks in this case study use earlier working names:
"Policy Score" for what shipped as the Driver Risk Score &
"Crash" for what shipped as Claims.
Test & Iterate
Usability Testing
Testing these designs with users surfaced needs already hinted at in earlier research, but prototypes confirmed they weren't just nice-to-haves. Without these features, we wouldn't have fixed this painful problem for customers.
Custom Incident Scoring
User Insight
Risk isn't one-size-fits-all. A ride-share company and an ambulance service don't weigh the same incident the same way.
Decision
Built custom rules settings, letting each company weight incident types by their unique safety policies, (i.e. an injury scores higher than property damage).
Incident score custom settings
Company settings page
Custom Risk Categories
User Insight
A single set of default risk categories (Clean, Activity, Exception) couldn't capture how differently companies needed to define "safe" vs "unsafe" drivers.
Decision
Introduced configurable risk score ranges per category, turning a raw number into an instant, at-a-glance signal.
Risk Score custom category settings
Adjustment Score
User Insight
Some of the clearest signals on a driver, such as telematics, training and positive performance, weren't captured by MVR or Claims at all.
Decision
Gave Safety Managers the ability to adjust an individual driver's score, case by case, directly from their profile. This score was added into the final Risk Score alongside MVR and Claims.
Adding an adjustment on a driver
Driver adjustments history
Final Driver Profile
Without all three scores working together, Samba would have solved nothing: Safety Managers would still be piecing together the same incomplete picture, just inside one platform instead of several. Combining MVR, Claims and Adjustment scores gave Safety Managers one number they could trust immediately, no cross-referencing required.
Final Driver Profile:
MVR + Claims + Adjustments scores = Risk Score
Design System
Components
New components for Claims and the Driver Risk Score were built inside Samba's existing library. Regular engineering conversations made sure every element was also written in code, not just Figma, so the library became a single source of truth. That discipline paid off in speed, consistency and longevity beyond this one project.
Outcome
Results
-
Samba went live with MVR, Claims and Adjustment scores unified under one Driver Risk Score.
-
This gave Safety Managers the complete view that 70% of surveyed customers said would be extremely valuable, and they began migrating claims data almost immediately. Sales and customer success were ready to sell and support before launch, with no gap between promise and delivery.
-
The foundation laid here went on to power Samba's next major release, the Risk Insights Dashboard, helping land the largest deal in company history and earning the confidence of Samba's leadership who had been closely invested from the start.
Team Impact
-
This engagement brought the Balanced Team model to Samba: designers, PMs and engineers working side by side, not in sequence. I led that effort on design, helping define hiring criteria and mentoring Samba's first in-house designer. My Crafted counterparts did the same with their own PM and engineering peers.
-
By launch, Crafted and Samba were operating as one integrated team, a foundation that carried straight into the Dashboard project. Watch Crafted explain the Balanced Team approach at Denver Startup Week.
Reflection
-
One of my favorite parts of this project was the client relationship itself. Samba's leadership trusted me as a designer and a true partner, not just design as a service.
-
Accessibility is one place I'd push harder if I could revisit this project. Relying on color alone to convey risk level was an existing feature, and while it works for most people, it excludes colorblind users entirely. Adding better labels, and patterns or icons, would have made the product easier to read for everyone, not just users with an impairment.
-
I'm proud of what this team built together, and just as excited for what came next.
Testimonials from Samba Safety
I had not worked with a UX Designer prior to the Crafted engagement. The designer brought a much-needed perspective to the team. She was always focused on the user, workflows, and validating our assumptions.
Product Manager
Working with the team at Crafted gave us a leg up through the definition and execution of some of our new and innovative products.
CTO
We added a new project on top of existing roadmap commitments, Crafted helped us quickly ramp up a Balanced Team to define the problem, frame the solution and execute on the concept in weeks instead of months.
EVP of Product