📌 Project: AI Cybersecurity Assistant
👤 Role: Principal Product Advisor · UX Lead · AI-first Strategist
🏢 Company: Dfend
📆 Timeline: Nov 2024 – Jun 2025
⚡ Focus: AI Security · Cybersecurity SaaS · Mobile-first UX
🛠 Services: UX Strategy · Product Design · Wireframing · Conversational UI
🎯 Impact: Designed an AI-driven chatbot for account protection, phishing defense, and passwordless workflows. Built wireframes and feature hierarchy directly with the CEO, creating scalable, user-friendly prototypes. Delivered a foundation for next-gen mobile identity security while streamlining complex cybersecurity concepts into intuitive, accessible flows..
The goal of Dfend was to design an AI assistant that proactively protects users from threats—acting before they even realize they’re at risk. A key feature was a threshold slider that let users choose how much autonomy to grant the AI, from “just notify me” to “handle it automatically.” Since this was a net-new product with no existing experience, the challenge was to establish trust while introducing a new paradigm: AI working on your behalf across privacy, banking, and security—solving problems before users knew they had them.
🧩 No Existing Baseline — A brand-new product meant designing the entire end-to-end experience from scratch, with no patterns to lean on.
🤖 AI Trust & Autonomy — Users needed to set comfort levels with AI taking actions on their behalf, from notifications to full autonomy.
🔒 Security-Critical Tasks — High-stakes actions (account recovery, suspicious login protection, password resets) required balancing friction with ease.
📱 Conversational UX — A chatbot-like interface had to handle complex branching logic in ways that felt natural and approachable.
👤 Wide User Spectrum — Needed to work for both tech-savvy users and those less familiar with digital security, ensuring inclusivity and adoption.
Our goal was to deliver seamless, proactive security that protected user accounts without constant intervention. Users wanted peace of mind, not ongoing alerts or friction. Leveraging advancements in large language models (LLMs), we envisioned an AI that could work quietly in the background—detecting risks and taking action before threats became disruptions.
To achieve this, we designed the AI to make intelligent, implied decisions based on activity monitoring, pattern recognition, and real-time data signals. This approach allowed the assistant to notify users when necessary, or act autonomously when trust thresholds were set, creating a smarter, more secure, and less intrusive experience.
Dfend was designed to go beyond traditional cybersecurity apps by offering a proactive, AI-driven solution. Instead of waiting for users to act, the system continuously monitors daily patterns, detects anomalies, and takes corrective action autonomously—notifying the user only when needed.
The design challenge was to make advanced security approachable for all levels of technical literacy. By keeping the interface conversational and intuitive, Dfend built trust without overwhelming users with jargon or complex workflows.
Looking forward, the design anticipates scalability through modular features: subscription models, advanced monitoring, and premium services can be layered in without compromising the core promise—peace of mind through invisible, intelligent protection.
The launch of Dfend introduced an entirely new category of AI-driven personal cybersecurity. By shifting from reactive alerts to proactive protection, the platform reduced user anxiety, simplified complex security tasks, and gave users greater confidence in their digital safety. The threshold slider for AI autonomy became a standout feature, empowering users to decide how much control to delegate. Early prototypes demonstrated that users completed more tasks, felt more secure, and engaged more consistently compared to traditional security tools.
🔒 Proactive Protection — Transformed security from reactive alerts into an AI assistant that neutralizes threats before disruption.
🎛 User-Controlled Autonomy — Introduced a threshold slider so users could choose between “just notify me” or “handle it for me.”
📉 Reduced Friction & Anxiety — Simplified complex flows (MFA, password resets, suspicious activity) into clear, conversational actions.
👥 Inclusive Adoption — Designed for all digital literacy levels, making cybersecurity approachable to a broader audience.
🧩 Future Scalability — Established a modular framework for subscription models, advanced monitoring, and premium services.
