Case study

Verikai: leading product design, brand, and marketing

Verikai is an AI-driven InsurTech company. Its platform gives insurance underwriters machine-learning risk scores for group health. I joined as Principal Product Designer, led product and brand design as Director of Design, then took over marketing as Senior Director of Design & Marketing.

Roles
Principal Product Designer, 2021. Director of Design, 2022–2023. Senior Director of Design & Marketing, 2023–2025.
Scope
Product design, design system, brand, website, content, marketing
Marketing results
+500% demo requests and +353% marketing engagement in one quarter
Company
Acquired by American Financial Group for ~$120M, effective December 2021
The Verikai platform: Acme Group's overview page in front of the all-groups list

01

One role across product, brand, and marketing

Verikai's customers are insurance underwriters. They use the platform to act on machine-learning risk scores, and the company's website described the product as comprehensive group health risk analysis.

I joined Verikai as Principal Product Designer in 2021 and was already doing the product design and branding. American Financial Group acquired the company that December, and Verikai kept operating as a standalone business. After the acquisition the Director of Design title formalized that role within the parent company's title structure: I set brand and product design strategy, ran design reviews, and guided engineers through the design work. From 2023 to 2025, as Senior Director of Design & Marketing, the role added marketing: I directed the brand overhaul, ran the marketing program, and built the Figma design system used across product and marketing.

The growth numbers on this page came from the marketing program. The product redesign has its own, smaller set of results.

Marketing raised demo requests 500% and marketing engagement 353% within one quarter, and LinkedIn engagement grew more than tenfold. On the product side, the redesigned CSV import cut import errors 90% and brought customer setup under 10 minutes.

02

What I inherited, and how the job grew

The platform

Underwriters had to act on risk scores they couldn't interrogate. The interface had grown around the model: inconsistent patterns, fragmented workflows, and predictions shown without the evidence and context a person needs to judge them.

Data import

Underwriters imported large CSV files, and the process was error-prone. Column headers rarely matched Verikai's fields, users abandoned uploads partway through, and support tickets piled up.

The brand

Verikai read as informal and lacking gravitas in the insurance industry. That undermined its credibility with the enterprise clients it was trying to win.

2021 · Principal Product Designer
My first role at Verikai. I was already doing the product design and branding. American Financial Group acquired Verikai that December.
2022–2023 · Director of Design
The title caught up with the work. It formalized the product and brand design role I already had, and placed it in the parent company's title structure. I set the design strategy for both, ran design reviews, kept stakeholders informed, and worked directly with the CEO.
2023–2025 · Senior Director of Design & Marketing
Marketing joined design under one role. The task was to reposition Verikai as a serious, trustworthy partner for risk assessment and underwriting decisions, and to make the website, collateral, and content say so consistently.

03

Product design: making the model's output legible

The redesign started from user research with underwriters and a map of their workflows. From there I established a unified design system and data visualization patterns that showed what a risk score drew on, where confidence was strong or weak, and what to check before acting on it.

I worked closely with engineering and data science so every pattern stayed faithful to what the model actually knew. The goal was output an underwriter could read, question, and defend to their own stakeholders.

Paper sketches of the all-groups list and the group view
Sketching the structure. Two views carry most of the work: a list of every group in an account, and a group view with that group's overview information. Both were laid out on paper first.
Mid-fidelity layout of the group listings page with a Create group button
Group listings, mid-fidelity. The navigation rail and the table, laid out without color or brand.
All groups: a dark navigation rail with collections, four headline numbers, and a table of groups showing stage, last activity, group size, and reports
All groups. One table for every group in the account. The rail holds create, search, and collections (open, in negotiation, closed, archive), so a list can be narrowed without leaving it. Each row shows where the group stands and which reports exist.
Report states
Report states. A report is ready, generating, or one click away. Generating resolves in place, so nobody has to leave the list to check on it.
Moving a group
Moving a group. Stage, watchers, and owner sit together at the top of the record. A change confirms itself and can be undone.
Group overview for Acme Group: a reports rail, group size, median age, and gender, the latest report, contact and address cards, and recent activity
Group overview. Reports, claims, census, and history share one rail. The key figures lead the page, and the latest report sits beside them.
Group information
Group information. Each figure carries its own picture: one dot per member, the age spread with the median marked, and the gender split. The census date sits in the corner, so the numbers always say where they came from.

The CSV import

I designed a column mapping interface that matches CSV columns to Verikai's fields automatically where it can, and clearly surfaces the columns that need a person's decision. Inline data previews let users check a mapping before committing, and distinct states mark successful matches, ignored columns, and fields that need attention.

The import went from a support-heavy process to a self-serve tool. Errors fell 90%, setup time dropped under 10 minutes, and cleaner inputs meant the model's predictions started from defensible ground.

Sketch of column matching: CSV columns mapped to fields, with matches checked and mismatches crossed
Column matching, sketched. After a file is uploaded, the first screen shows the initial matching analysis: each column either matched or flagged.
Flow diagram of the upload CSV flow: upload, then column mapping
The flow. Upload, then mapping: the system matches columns with identical names, and the user confirms or picks the right field for the rest.
Column mapping
Column mapping. Columns with matching names map themselves. Anything the system can't place is flagged and opens a field picker, and ignored columns stay visible, so nothing disappears silently. A sample value beside each column lets people check a match before they commit.
Census upload
Census upload. A new file is checked before it replaces the current census, and each check reports as it passes.
Census page for Acme Group: an upload area, members over time, and a table of uploaded census files with who uploaded each one
Census history. Every file the group has sent, who uploaded it, and which reports used it.

04

Taking over marketing

When marketing joined my role, the problem in front of it was credibility with enterprise insurance buyers. The work ran in phases: agree on who Verikai was, then rebuild everything that carried the message.

  1. Starting problem

    A brand that undercut the sale

    Verikai was perceived as informal and lacking gravitas, which undermined credibility with the enterprise clients it needed.

  2. Intervention

    Reposition, with the executives on board

    I assembled the key stakeholders and established brand guidelines, with executive buy-in secured at an off-site retreat. The position: a serious, trustworthy partner for risk assessment and underwriting decisions.

  3. Execution

    Rebuild every surface

    • A complete website redesign
    • All marketing collateral updated
    • A content roadmap: daily LinkedIn posts, monthly newsletters, and outreach to industry publications
    • An AI-driven marketing program
  4. Measured result

    +500%

    demo requests within one quarter

    +353%

    marketing engagement in the same quarter

    10×+

    LinkedIn engagement

Redesigned Verikai website: 'Minimize risk. Maximize accuracy.' hero and a group health risk analysis section
The redesigned website. The new position in one line, "Minimize risk. Maximize accuracy.", over the product it describes: comprehensive group health risk analysis.
Verikai insights newsletter and collateral pages in the new visual system
Newsletter and collateral. The monthly "Verikai insights" issue and supporting pieces, rebuilt in the new system.
Verikai print pieces and an industry magazine placement
Print and industry publications. Printed pieces in the new system alongside an industry magazine, from the publication outreach.

05

One story and one visual system

I built a single Figma design system and used it for both the product and marketing. The same deep indigo runs through the platform's navigation, the website, the newsletter, and the sales materials, so the product and every marketing surface read as one company.

Brand guidelines page: Mulish display font and the color palette, primary #25378C
The foundation. Mulish for display, a primary indigo (#25378C), and a small set of accents with defined jobs, such as links and buttons.
Grid of Verikai presentation slides with charts in the brand palette
Presentation slides. Data slides with charts, drawn in the brand palette.
Grid of Verikai editorial content covers in the brand's indigo
Editorial covers. One photograph and one headline per piece, set in the brand indigo.

06

What I owned, did, and led

Owned

Product design and branding from 2021, with the strategy formally mine as Director. Design and marketing together from 2023.

Did myself

Built the Figma design system used across product and marketing. Designed the CSV column mapping interface. Led the platform redesign from research through to the final screens.

Led and partnered

Ran design reviews and guided engineers through the design work. Kept stakeholders informed and worked directly with the CEO. Partnered with engineering and data science on the platform, and with company leadership on the repositioning. Directed the brand overhaul and the marketing program.

07

Results, and where each one came from

Marketing

+500%

demo requests within one quarter

+353%

marketing engagement in the same quarter

10×+

LinkedIn engagement

Product

−90%

CSV import errors after the mapping redesign

<10 min

customer setup time

Company context

~$120M

American Financial Group's acquisition of Verikai, effective December 2021, early in my time there. The leadership and marketing work on this page came after it.

Lessons

What leading across product and marketing taught me

  1. Make the model legible before making it pretty. The redesign showed what a score drew on and how confident it was, so underwriters could question a prediction and defend it.
  2. Fix the inputs, not just the outputs. The CSV import looked like a support problem. It was also a model problem: cleaner inputs meant predictions started from defensible ground.
  3. Credibility comes before content. Guidelines and executive agreement came first. The website, collateral, and content roadmap were built on them, and that's where the quarter's growth showed up.

Where this led

How should a system show its evidence so people calibrate their trust correctly?

Verikai was the clearest early version of that question. SlateRoom is where I'm working on it now.

Read the SlateRoom case study

Contact

If you're shaping an ambitious product and need someone who can set its direction, design it, and help build it, let's talk.

carl.brunson@gmail.com