We are committed to increasing diversity in our workplace and supporting social justice in our communities.
We believe that the professional is personal and the personal is professional, and we strive to foster an environment that welcomes open and honest discourse about what it means to be truly equal. To us these aren’t just buzzwords that stand for the tolerance of differences, but the radical inclusion of people from all walks of life so that everyone has a seat at the table.
Diverse Teams Build Diverse Models
There is no shortage of research highlighting the dangers of implicit bias in AI, that decision-making tools created by homogenous teams disparately impact non-dominant groups. Since humans create the models that power our automation and operational efficiency solutions for some of the world’s largest insurers, we understand the critical importance of diversifying our data science and research teams, as well as building safeguards in the development process to ensure that each new data element can be applied ubiquitously and without bias.
Carpe Data helps bridge the gap between legacy risk assessment and changing regulatory standards, and one of our founding missions is to replace discriminatory data practices with equally predictive, next-generation alternatives that make insurance claims assessment and underwriting more equitable for all.
In Claims Assessment
Carpe Data’s ClaimsX helps claims teams more effectively leverage online and social data to expedite low-risk claims or dive deeper into more complex ones. Historically, claims handlers might manually “google” claimants or crawl their social media presence for corroborative or contradictory evidence—not only a time-consuming practice but one rife with opportunities for bias, implicit or otherwise. By automating online searches and only reporting information relevant to the claim, ClaimsX removes subjective analysis or “gut feelings” from the investigation process, protecting consumers as well as insurers.
In Business Underwriting
It takes a lot of data to truly understand the potential risks that business owners face when they start a new venture or expand their current operations. Traditional data sources often leverage overly simplistic elements like zip code or even crime rates to assess risk—elements that paint entire business communities (often of color) with broad strokes and do a poor job of assessing an entity’s actual business performance, resulting in higher premiums regardless of business operations.
Our Minerva business data suite seeks to upend these elements with new rating factors—like online visibility, reputation, and sanitary conditions—that better assess individual business risk without misclassifying or stereotyping entire neighborhoods. These (and many other) new data points combine to provide more accurate risk profiles than traditional sources ever did; it’s a win-win for growing business communities and insurers alike.
Accountability Through Data
As a data company we know that the best way to effect change is to back it up with hard figures. By focusing on core human metrics like recruitment, payroll, performance management and professional development, we uncover patterns of systemic bias within our own systems and practices, ultimately allowing us to focus resources on those hotspots. We hold ourselves accountable to these data through key internal barometers, including:
Two-way sentiment feedback holds us accountable to each other—from entry level to executive leadership, all feedback is considered and evaluated equally.
Quarterly demographic updates, both internally and externally, keep a public record of our progress in creating a company culture that better reflects our society.
Who We Are: By the Numbers
DEI in Hiring and Promotion Practices
From recruitment sourcing to final offers, we’re reinventing our hiring practices to eliminate bias. Learn more about each stage of our interview process on our Careers page.
Promoting from Within
A 2021 article by Harvard Business Review revealed that where there are at least two non-dominant candidates in the finalist pool for a promotion (regardless of the size of the candidate pool), the chances that a non-dominant candidate will get the job multiplied by a factor 193. Creating a diverse promotion pipeline significantly impacts organizational depth and growth, as well as continuity and increased morale.
Diversifying Hiring Pipelines
Growing companies like ours tend to look a lot like the industries they serve. To combat implicit bias often found while hiring for “industry expertise”, we have committed to using new and non-traditional recruiting pipelines for new hires and educating them on insurance—the oldest data science industry in the world—along the way.
Company Initiatives & Employee Resource Groups:
This group meets monthly and fosters an intentional space where female-identifying team members and allies discuss current topics around what it means to be a woman, both in and outside of work.
PORTUGUESE WOMEN IN TECH
Portuguese Women in Tech was launched in April 2016 with a dual mission: to support women in technology by providing visibility, networking, mentorship and trainings; and to attract more women and girls to tech and in this way, diversify traditionally male hiring pipelines.
Our company-wide book club collaborates on book choice, focused on authors of color and topics of social justice. Previous discussion books include: Biased, Caste; The Origins of our Discontents, We Should All Be Feminists, and The Likeability Trap.
Carpe Data also provides access to UCSB Arts & Lectures, which features a vast array of artists, scientists, authors, and thinkers from around the world.
- Recent Team Workshops:
- Company-wide implicit bias training
- Taking Pride in(to) Your Work: How to foster inclusive environments for LGBTQIA+ employees
- Gather & Learn: an internal program for employees to share topics that they’re passionate about. Past topics include:
- How to support refugee communities in crisis
- Cross-cultural understanding between US and Portugal teams
- Recognizing microaggressions