Making thousands of signals feel like a few decisions

Making thousands of signals feel like a few decisions

Correlates vulnerabilities, identities, misconfigurations, and live threat signals — condensing fragmented alerts from 12+ tools into a handful of prioritized exposure decisions.

Correlates vulnerabilities, identities, misconfigurations, and live threat signals — condensing fragmented alerts from 12+ tools into a handful of prioritized exposure decisions.

Sentra AI is an enterprise cybersecurity platform that correlates vulnerabilities, identities, and threat signals into prioritized exposure decisions. Instead of reviewing thousands of isolated alerts, analysts investigate prioritized risks with business context.

Sentra AI is an enterprise cybersecurity platform that correlates vulnerabilities, identities, and threat signals into prioritized exposure decisions. Instead of reviewing thousands of isolated alerts, analysts investigate prioritized risks with business context.

Problem

Problem

For security analysts, the problem isn't a shortage of data it's a shortage of clarity. Thousands of alerts. No shared context. No way to know what actually matters.

For security analysts, the problem isn't a shortage of data it's a shortage of clarity. Thousands of alerts. No shared context. No way to know what actually matters.

My first instinct was to make it simpler. I was wrong.

My first instinct was to make it simpler. I was wrong.

When I first mapped out all the data that needed to live on screen, I did what most designers would do — I tried to reduce it. Hide things. Create summary views. Show less upfront.

When I first mapped out all the data that needed to live on screen, I did what most designers would do — I tried to reduce it. Hide things. Create summary views. Show less upfront.

But after spending time understanding who actually uses this — security analysts, CISOs, SecOps leads — I realised the problem wasn't too much information. It was information with no hierarchy. No way to know what mattered most.

But after spending time understanding who actually uses this — security analysts, CISOs, SecOps leads — I realised the problem wasn't too much information. It was information with no hierarchy. No way to know what mattered most.

These users work long sessions. They need density. They need to scan a table and know in 2 seconds what's critical. The design challenge wasn't subtraction — it was hierarchy, contrast, and clarity at speed.

These users work long sessions. They need density. They need to scan a table and know in 2 seconds what's critical. The design challenge wasn't subtraction — it was hierarchy, contrast, and clarity at speed.

My role & what I was working within

My role & what I was working within

TIMELINE

2 Months

MY ROLE

Product Designer

TEAM

2 Designers

I owned UI design across all four sections of the platform — Dashboard, Exposures, Landscape, and the AI co-pilot. I wasn't leading product strategy or conducting user research. Every visual and interaction decision that shaped how users experienced the product was mine.

I owned UI design across all four sections of the platform — Dashboard, Exposures, Landscape, and the AI co-pilot. I wasn't leading product strategy or conducting user research. Every visual and interaction decision that shaped how users experienced the product was mine.

THE NON-NEGOTIABLE CONSTRAINT

Designing for domain experts under pressure

Security analysts don't have time to learn a new interface. Every pattern had to be instantly readable. Density was a requirement, not a problem. Familiarity was a feature. That constraint shaped every single layout decision I made across the product.

Security analysts don't have time to learn a new interface. Every pattern had to be instantly readable. Density was a requirement, not a problem. Familiarity was a feature. That constraint shaped every single layout decision I made across the product.

Not what I designed, how I got there

Not what I designed, how I got there

Each decision below started from a version that didn't work. The story of how I got from that first version to what shipped is where the actual design thinking lives.

Each decision below started from a version that didn't work. The story of how I got from that first version to what shipped is where the actual design thinking lives.

DECISION 01

Progressive disclosure over full context upfront

My first version of the exposure detail page showed everything — narrative, evidence, attack path, TTPs, impact — all on one scrolling page. It was complete. It was also overwhelming. A CISO looking at this at 9am doesn't need everything. They need the score and the one thing that matters most. I rebuilt it as four tabs. Each level earns the next.

My first version of the exposure detail page showed everything — narrative, evidence, attack path, TTPs, impact — all on one scrolling page. It was complete. It was also overwhelming. A CISO looking at this at 9am doesn't need everything. They need the score and the one thing that matters most. I rebuilt it as four tabs. Each level earns the next.

DECISION 02

The risk simulator: making remediation feel actionable

Security teams know their risk score is 95. They don't know which 2 actions bring it to 70. I explored showing a static recommendations list — but it still required mental math. The slider changed everything. Set the score you want. See which actions get you there, with effort level and score delta, before you commit to anything.

Security teams know their risk score is 95. They don't know which 2 actions bring it to 70. I explored showing a static recommendations list — but it still required mental math. The slider changed everything. Set the score you want. See which actions get you there, with effort level and score delta, before you commit to anything.

DECISION 03

Proactive AI: open with a brief, not a blank prompt

The first version of Ask Osfi was just a chat interface. Clean, minimal, exactly what you'd expect. And it asked the user to do all the work. For a security analyst at 8am who doesn't know what to ask — a blank prompt is friction, not help. I designed it to open with a proactive security brief. The user gets value before typing a single character.

please use web view for more

The Screens

Each decision below started from a version that didn't work. The story of how I got from that first version to what shipped is where the actual design thinking lives.

Contact

I’m always excited to collaborate on new work. Let’s create something great together!

Copy Email

Create a free website with Framer, the website builder loved by startups, designers and agencies.